Download the full version of the ebook now at ebookultra.

com

Biotechnology Annual Review Vol 11 1st

Edition M.R. El-Gewely (Ed.)

https://ebookultra.com/download/biotechnology-
annual-review-vol-11-1st-edition-m-r-el-gewely-ed/

Explore and download more ebook at https://ebookultra.com

 Recommended digital products (PDF, EPUB, MOBI) that
you can download immediately if you are interested.

Biotechnology Annual Review 13 1st Edition M. Raafat El-

Gewely (Eds.)

https://ebookultra.com/download/biotechnology-annual-review-13-1st-
edition-m-raafat-el-gewely-eds/

ebookultra.com

Annual Review of Psychology vol 54 2003 Susan T. Fiske

https://ebookultra.com/download/annual-review-of-psychology-
vol-54-2003-susan-t-fiske/

ebookultra.com

Annual Review of Psychology Vol 57 2006 1st Edition Susan

T. Fiske

https://ebookultra.com/download/annual-review-of-psychology-
vol-57-2006-1st-edition-susan-t-fiske/

ebookultra.com

Annual Review of Psychology Vol 58 2007 1st Edition Susan

T. Fiske

https://ebookultra.com/download/annual-review-of-psychology-
vol-58-2007-1st-edition-susan-t-fiske/

ebookultra.com
Annual Review of Pharmacology and Toxicology Volume 45
2005 Annual Review of Pharmacology and Toxicology 1st
Edition Arthur K. Cho
https://ebookultra.com/download/annual-review-of-pharmacology-and-
toxicology-volume-45-2005-annual-review-of-pharmacology-and-
toxicology-1st-edition-arthur-k-cho/
ebookultra.com

Advances in Insect Physiology Vol 29 1st Edition Peter

Evans

https://ebookultra.com/download/advances-in-insect-physiology-
vol-29-1st-edition-peter-evans/

ebookultra.com

Demon Slayer Kimetsu no Yaiba Vol 11 11 Gotouge

https://ebookultra.com/download/demon-slayer-kimetsu-no-yaiba-
vol-11-11-gotouge/

ebookultra.com

Annual Review of Scalable Computing 1st Edition Yuen Chung

Kwong

https://ebookultra.com/download/annual-review-of-scalable-
computing-1st-edition-yuen-chung-kwong/

ebookultra.com

Oceanography and Marine Biology An Annual Review Volume 41

An Annual Review Volume 41 Oceanography and Marine Biology
1st Edition R. N. Gibson
https://ebookultra.com/download/oceanography-and-marine-biology-an-
annual-review-volume-41-an-annual-review-volume-41-oceanography-and-
marine-biology-1st-edition-r-n-gibson/
ebookultra.com
Biotechnology Annual Review Vol 11 1st Edition M.R. El-
Gewely (Ed.) Digital Instant Download
Author(s): M.R. El-Gewely (Ed.)
ISBN(s): 9780444519528, 0444519521
Edition: 1st
File Details: PDF, 9.11 MB
Year: 2005
Language: english
 v

Preface

Biotechnology is a diverse, complex and rapidly evolving field. Students and

experienced researchers alike face the challenges of staying on top of developments
in their field of specialty and maintaining a broader overview of the field as a
whole. Volumes containing competent reviews on a diverse range of topics in the field
fulfill the dual role of broadening and updating biotechnologists’ knowledge. The
current volume, edited by Dr. M. Raafat El-Gewely is an excellent example of such a
book. The topics covered range from classical issues in biotechnology – such as,
vehicles for the production of biotechnology products and methods for their
detection, separation and analysis – to topics that are focused on the role of
biotechnology in the health sciences. The information presented in this book will
therefore will be of great value to both experienced biotechnologists and biotechnol-
ogists in training.
In a field that is as dynamically changing as the one covered in this volume,
it may be tempting to ask what topics the chapters of the Biotechnology Annual
Review might include in the future. Two directions which are very likely to gain
greater prominence are the integration of diverse types of quantitative data for
the generation of models that represent the structure and dynamics of cellular
pathways; and the use of new platforms and detection systems developed in
biotechnology for the detection of diagnostic and prognostic markers or patterns
of markers.
The first direction is at the heart of the emerging science of systems biology.
Considering the rapid advances in technologies and platforms for the collection
of large and quantitative datasets on gene expression, proteome and metabolite
profiles and interaction maps among proteins and between proteins and nucleic
acids, it must be expected that the demand for advanced computational tools for
the analysis, validation, integration and display of the data will dramatically
increase. The latter direction, if successful, represents the type of research that
has the potential for a tremendous impact on public health through the earlier
and more precise diagnosis of diseases as well as for the development of new
drugs and the optimized use of existing drugs. This volume reviews both of
these promising pathways.
By demonstrating a commitment to comprehensive coverage of issues which
constitute the current field of biotechnology and a willingness to anticipate those
vi

which may shape its future, Dr. El-Gewely reinforces the role of the Biotechnology
Annual Review as a stimulating and informative resource for researchers now and
in years to come.

Ruedi Aebersold, Ph.D.

Professor of Systems Biology
Institute of Biotechnology
Swiss Federal Institute of Technology
ETH Hönggerberg HPT E 78
CH-8093 Zürich
Phone: 41 1 633 31 70; FAX: 41 1 633 10 51
[email protected]

And:
Professor of Systems Biology
The Institute for Systems Biology
1441 North 34th Street
Seattle, WA 98103-8904, USA
Phone: 1-(206) 732-1204; Fax: 1-(206) 732-1299
[email protected]
http://www.systemsbiology.org/
 vii

Contributions

Potential contributors/authors of an appropriate chapter wishing to publish it in

Biotechnology Annual Review should contact the Chief Editor, or any member of the
Editorial Board as listed. For the ‘Guide for Authors’, visit: www.elsevier.com/
locate/issn/13872656
 ix

Editorial Board

Chief editor Clayton, Victoria 3800

Tel +61-3-9905 3711
Dr. M. Raafat El-Gewely Fax +61-3-9905 4655
Department of Molecular Biotechnology E-mail: Alfons.Lawen@med.
Institute of Medical Biology monash.edu.au
University of Tromsø
9037 Tromsø, Norway Associate editors
Phone: 47-77 64 46 54
Fax: 47-77 64 6290 Dr. Marin Berovic
E-mail: [email protected] Department of Chemical and
Biochemical Engineering
University of Ljubljana
Editors Hajdrihova 19
Ljubljana
Dr. MaryAnn Foote, Solvania
Director, Medical Writing Amgen Inc. E-mail: [email protected]
One Amgen Center Drive, M/S 24-1-C
Thousand Oaks, CA 91320-1799 Dr. Thomas M.S. Chang
USA Artificial Cells & Organs Research
Phone: 1-805-447-4925 Centre
Fax: 1-805-498-5593 McGill University
E-mail: [email protected] 3655 Drummond St., Room 1005
Montreal, Quebec, Canada H3G 1Y6
Dr. Guido Krupp Phone: 1-514-398-3512
Director & Founder Fax: 1-514-398-4983
artus GmbH E-mail: [email protected]
Koenigstr. 4a
D-22767 Hamburg, Germany Dr. Thomas T. Chen
Phone: +49 - 40 - 41 364 783 Department Molecular and Cellular
Fax: +49 - 40 - 41 364 720 Biology
E-mail: [email protected] University of Connecticut
Website: www.artus-biotech.com 91 North Eagleville Rd,
Unit 3125
Dr. Alfons Lawen Storrs,
Senior Lecturer Connecticut 06269-3149,
Monash University, Clayton Campus USA
Department of Biochemistry and Phone: 1-860- 486-5481
Molecular Biology Fax: 1-860- 486-5005
Room 312, Building 13D, E-mail: [email protected]
x

Dr. Frank Desiere Phone: 81-75-753 6266

Nestlé Research Center, Fax: 81-75-753 6265
P.O. Box 44, E-mail: [email protected]
CH-1000 Lausanne 26
Switzerland Dr. Jocelyn H. Ng
E-mail: [email protected] 2/6 Jurang St.
Balwyn VIC 3103
Professor Franco Felici Australia
Department of Microbiological, Phone: +61 3 9836 1679
Genetic and Molecular Science, Fax: +61 3 9836 1699
University of Messina, E-mail: [email protected]
Salita Sperone 31, 98166 Messina
Italy Dr. Eric Olson
Phone: +39 0906765197 Vertex Pharmaceuticals
Fax: +39 090392733 130 Waverly Street
E-mail: [email protected] Cambridge, MA 02139
USA
Dr. Leodevico (Vic) L. Ilag Phone: 1-858-404-5381
Chief Scientific Officer Fax: 1-858-404-6787
Cryptome Pharmceuticals Ltd E-mail: [email protected]
Level 1, Baker Heart Research Institute
Bldg Dr. Steffen B. Petersen
Commercial Road Biostructure and Protein Engineering
Melbourne VIC 3004 Laboratory
Australia Department of Biotechnology
PO Box 6492 University of Aalborg
St Kilda Road Central Sohngaardsholmsvej 57
Melbourne VIC 8008 DK-9000 Aalborg
Phone: +61 3 8532 1725 Denmark
Fax: + 61 3 8532 1721 Phone: 45-9-635 8469
E-mail: Fax: 45-9-814 2555
[email protected] E-mail: [email protected]
www.cryptomepharmaceuticals.com
Prof. Vincenzo Romano-Spica
Professor Kuniyo Inouye, Ph.D Professor of Hygiene
Laboratory of Enzyme Chemistry University Institute of Motor Science,
Division of Food Science and IUSM
Biotechnology P.zza Lauro e Bosis 15, 00194 Rome,
Graduate School of Agriculture Italy
Kyoto University Phone/Fax: +39-06-36733247
Sakyo-ku, Kyoto 606-8502, Japan E-mail: [email protected]
 xi

List of contributors

Katsuo Aizawa Phone: +39 059 2055474

Department of Physiology Fax: +39 059 2055483
Tokyo Medical University E-mail: [email protected]
Tokyo 160-8402, Japan
Mohamed Boudjelal
Karen Arts Department of Gene Expression and
Amgen Canada Inc Protein Biochemistry
Missisauga, Canada GlaxoSmithKline
Discovery Research, Harlow, UK
Marin Berovic
Department of Chemical, Biochemical Joseph Brandwein
& Environmental Engineering University Health Network
Faculty of Chemistry and Toronto, ON, Canada
Chemical Engineering
University of Ljubljana Carol Box
Askerceva 5, 1000 Ljubljana, Slovenia Tumour Biology and Metastasis,
Phone: +38 61 249510 Cancer Research UK
Fax: +38 61 4760300 Centre for Cancer Therapeutics
E-mail: marin.berovic@f kkt.uni-lj.si McElwain Laboratories
Institute of Cancer Research
Michael V. Berridge Cotswold Road, Belmont, Sutton
Malaghan Institute of Medical Research Surrey, SM2 5NG, UK
PO Box 7060, Wellington
New Zealand William Cairns
Fax: 0064 4 499 6915. Department of Gene Expression and
E-mail: [email protected] Protein Biochemistry
GlaxoSmithKline Discovery Research
Moreno Bondi Harlow, UK
Department of Biomedical Sciences Phone: +44 1279 622043
University of Modena and Fax: +44 1279 627666
Reggio Emilia E-mail: [email protected]
Via Campi 287, 411 000 Modena,
Italy O. Cattarini
University of Movement Sciences
Paola Borella (IUSM)–Section of Hygiene
Department of Hygiene and De Bosis 6-00194
Microbiology Foro Italico, Rome, Italy
University of Modena and
Reggio Emilia J. Patrick Condreay
Via Campi 287, 41100 Modena, Department of Gene Expression and
Italy Protein Biochemistry
xii

Research Triangle Park Franco Felici

NC, USA Department of Microbiological,
Genetic and Molecular Science
University of Messina, Italy
William Court
Tumour Biology and Metastasis
Hans Peter Fischer
Cancer Research UK
Genedata AG, Basel, Switzerland
Centre for Cancer Therapeutics
E-mail: Hans-Peter.fischer@
McElwain Laboratories
genedata.com
Institute of Cancer Research
Cotswold Road, Belmont, Sutton
Jonathan M. Fleming
Surrey, SM2 5NG, UK
Department of Gene Expression and
Protein Biochemistry
Michael Crump
GlaxoSmithKline Discovery Research
Amgen Inc., Thousand Oaks
Harlow, UK
California, USA
MaryAnn Foote
P. Di Michele
Amgen Inc., Thousand Oaks
University of Movement Sciences
California, USA
(IUSM)–Section of Hygiene
De Bosis 6-00194
Edmee Franssen
Foro Italico, Rome, Italy
Toronto-Sunnybrook Regional
Cancer Ctr
Dale A Dotten
Toronto, ON, Canada
Michael’s Hospital
Toronto, ON, Canada
Gabriella Garufi
Suzanne A. Eccles Department of Microbiological,
Tumour Biology and Metastasis Genetic and Molecular Science
Cancer Research UK University of Messina, Italy
Centre for Cancer Therapeutics
McElwain Laboratories Bruce German
Institute of Cancer Research University of California in Davies
Cotswold Road, Belmont, CA, USA
Sutton, Surrey
SM2 5NG, UK G. Gianfranceschi
Phone: + 44 20 87224210 University of Movement Sciences
Fax: + 44 20 87224134 (IUSM)–Section of Hygiene
E-mail: [email protected] De Bosis 6-00194
Foro Italico, Rome, Italy
Natalja Skrebova Eikje
Tartu Ulikooli Nahahaiguste Kliinik Elisa Guerrieri
Raja 31, 50417 Tartu, Estonia Department of Biomedical Sciences
Phone: + 37 24 834 16 University of Modena and
Fax: + 47 52 7777 Reggio Emilia
E-mail: [email protected] Via Campi 287, 411 00 Modena, Italy
 xiii

Patries M. Herst E. Montuori

Malaghan Institute of Medical Research University of Movement Sciences
PO Box 7060, (IUSM)–Section of Hygiene
Wellington, New Zealand De Bosis 6-00194
Foro Italico, Rome, Italy
Roy M. Katso
Assay Development and M. Orsini
Compound Profiling University of Movement Sciences
Stevenage, UK (IUSM)–Section of Hygiene
De Bosis 6-00194
Isabella Marchesi Foro Italico, Rome, Italy
Department of Hygiene and
Microbiology Yukihiro Ozaki
University of Modena and Department of Chemistry
Reggio Emilia School of Science and Technology
Via Campi 287, 411 00 Modena, Italy Kwansei-Gakuin University
Sanda 669-1337, Japan
Sarah J. Mason
Department of Gene Expression and A. Paparini
Protein Biochemistry University of Movement Sciences
GlaxoSmithKline Discovery Research (IUSM)–Section of Hygiene
Harlow, UK De Bosis 6-00194
Foro Italico, Rome, Italy
Patrizia Messi
Department of Biomedical Sciences Janet H. Parham
University of Modena and Department of Gene Expression and
Reggio Emilia Protein Biochemistry
Via Campi 287, 411 00 Modena, Italy Research Triangle Park NC, USA

Raymond V. Merrihew Carla Lo Passo

Assay Development and Compound Department of Microbiological,
Profiling Genetic and Molecular Science
Research Triangle Park, University of Messina, Italy
NC, USA
Ida Pernice
Olga Minenkova Department of Microbiological,
Kenton Srl, c/o Sigma-Tau, Pomezia Genetic and Molecular Science
Rome, Italy University of Messina, Italy

Monica Monici Aleš Podgornik

CEO Centre of Excellence in Optronics, BIA Separations d.o.o.
Florence, Italy Teslova 30, 1000 Ljubljana, Slovenia
E-mail: [email protected] Phone: +386 1 426 56 49
Phone: 39-055-4271217 Fax: +386 1 426 56 50
Fax: 39-055-2337755 E-mail: [email protected]
xiv

V. Romano-Spica Aleš Štrancar

University of Movement Sciences BIA Separations d.o.o.,
(IUSM)–Section of Hygiene Teslova 30
De Bosis 6-00194 SI-1000 Ljubljana, Slovenia
Foro Italico, Rome, Italy
Phone/Fax: +39-(06)-36733247 An S. Tan
E-mail: [email protected] Malaghan Institute of Medical Research
PO Box 7060, Wellington
Jean-Pierre G Routy New Zealand
Royal Victoria Hospital
Montreal, Canada Marianne Taylor
Cancer Centre for Southern Interior
Carol Anne Sawka Kelowna, BC, Canada
Toronto-Sunnybrook Regional
Cancer Ctr G.C. Vanini
2075 Bayview Ave University of Movement Sciences
Toronto, ON M4N 3M5 Canada (IUSM)–Section of Hygiene
Phone: 416 217 1282 De Bosis 6-00194
Fax: 416 971 6888 Foro Italico, Rome, Italy
E-mail: [email protected]
O. Vassioukovitch
Frances A Shepherd University of Movement Sciences
University of Toronto (IUSM)–Section of Hygiene
Toronto, ON, Canada De Bosis 6-00194
Foro Italico, Rome, Italy
Anne-Marie Stomp
of Forestry Irwin R Walker
North Carolina State University Hamilton Health Sciences Corp
Raleigh, NC 27695-8002 Hamilton, ON
E-mail: [email protected] Canada

Jean St-Louis Heribert Watzke

Hopital Maisonneuve-Rosemont Nestlé Research Center, Lausanne
Montreal, Canada Switzerland
 xv

Contents
Preface v
Editorial Board ix
List of contributors xi

Towards quantitative biology: Integration of biological information

to elucidate disease pathways and to guide drug discovery
Hans Peter Fischer 1
The duckweeds: A valuable plant for biomanufacturing
Anne-Marie Stomp 69
The application of BacMam technology in nuclear receptor drug discovery
Mohamed Boudjelal, Sarah J. Mason, Roy M. Katso,
Jonathan M. Fleming, Janet H. Parham, J. Patrick Condreay,
Raymond V. Merrihew and William J. Cairns 101
Tetrazolium dyes as tools in cell biology: New insights into their
cellular reduction
Michael V. Berridge, Patries M. Herst and An S. Tan 127
Display libraries on bacteriophage lambda capsid
Gabriella Garufi, Olga Minenkova, Carla Lo Passo, Ida Pernice
and Franco Felici 153
Vibrational spectroscopy for molecular characterisation and
diagnosis of benign, premalignant and malignant skin tumours
Natalja Skrebova Eikje, Katsuo Aizawa and Yukihiro Ozaki 191
Cell and tissue autofluorescence research and diagnostic applications
Monica Monici 227
Sterilisation in biotechnology
Marin Berovic 257
Õ
Convective Interaction Media (CIM) – Short layer monolithic
chromatographic stationary phases
Alesˇ Podgornik and Alesˇ Sˇtrancar 281
Detection of metazoan species as a public health issue: simple
methods for the validation of food safety and quality
O. Vassioukovitch, M. Orsini, A. Paparini, G. Gianfranceschi,
O. Cattarini, P. Di Michele, E. Montuori, G.C. Vanini
and V. Romano Spica 335
Water ecology of Legionella and protozoan: environmental and
public health perspectives
Paola Borella, Elisa Guerrieri, Isabella Marchesi, Moreno Bondi
and Patrizia Messi 355
xvi

A Prospective, non-randomised phase 1–2 trial of VACOP-B

with filgrastim support for HIV-related non-Hodgkin’s lymphoma
Carol Anne Sawka, Frances A. Shepherd, Edmee Franssen,
Joseph Brandwein, Dale A. Dotten, Jean-Pierre G. Routy,
Irwin R. Walker, Jean St-Louis, Marianne Taylor, Karen Arts,
Michael Crump and MaryAnn Foote 381
Cell migration/invasion assays and their application in
cancer drug discovery
Suzanne A. Eccles, Carol Box and William Court 391

Index of authors 423

Keyword index 425

 1

Towards quantitative biology: Integration of biological

information to elucidate disease pathways and to guide drug
discovery
Hans Peter Fischer*
Genedata AG, Basel, Switzerland

Abstract. Developing a new drug is a tedious and expensive undertaking. The recently developed
high-throughput experimental technologies, summarised by the terms genomics, transcriptomics,
proteomics and metabolomics provide for the first time ever the means to comprehensively monitor
the molecular level of disease processes. The ‘‘-omics’’ technologies facilitate the systematic
characterisation of a drug target’s physiology, thereby helping to reduce the typically high attrition
rates in discovery projects, and improving the overall efficiency of pharmaceutical research
processes. Currently, the bottleneck for taking full advantage of the new experimental technologies
are the rapidly growing volumes of automatically produced biological data. A lack of scalable
database systems and computational tools for target discovery has been recognised as a major
hurdle. In this review, an overview will be given on recent progress in computational biology that
has an impact on drug discovery applications. The focus will be on novel in silico methods to
reconstruct regulatory networks, signalling cascades, and metabolic pathways, with an emphasis
on comparative genomics and microarray-based approaches. Promising methods, such as the
mathematical simulation of pathway dynamics are discussed in the context of applications in
discovery projects. The review concludes by exemplifying concrete data-driven studies in
pharmaceutical research that demonstrate the value of integrated computational systems for
drug target identification and validation, screening assay development, as well as drug candidate
efficacy and toxicity evaluations.

Keywords: systems biology, drug discovery, functional genomics, data integration, genomics,
transcriptomics, proteomics, metabolomics, microarray data, whole-genome sequences, gene
prediction, data analysis, genome analysis, sequence analysis, expression analysis, expression
profiling, tissue expression, tissue profiling, target identification, target validation, assay
development, reporter assays, compound evaluation, oncogenomics, biomedical research, disease
hypothesis, diagnostics, marker genes, cancer research, oncology, antibiotics, anti-microbial,
chemotherapy, in silico, efficacy studies, toxicity studies, pathway simulations, pathway dynamics,
pathway modelling, quantitative biology, computational biology, bioinformatics, -omics
technologies, research management system, drug development, compound screening, high-
throughput screening, mode-of-action, mechanism-of-action, MOA, promoter analysis, promoter
modules, regulatory elements, composite elements, gene regulation, position weight matrix, TFBS,
metabolic pathways, regulatory pathways, pathway reconstruction, regulatory modules, regulatory
network, signalling pathways, phylogenetic footprinting, transcription factor binding site
prediction, toxicogenomics, pharmacogenomics, fusion protein analysis, phylogenetic profiling,
genome comparisons, comparative genomics, COG, gene neighbourhood analysis, operon
prediction, computational methods, yeast-two-hybrid, protein interactions, 2D-gels, oligonucleotide
chips, DNA chips, mass spectrometry, LC/MS, GC/MS, MS/MS, metabolic flux, fluxomics.

*E-mail: [email protected]

BIOTECHNOLOGY ANNUAL REVIEW ß 2005 ELSEVIER B.V.

VOLUME 11 ISSN: 1387-2656 ALL RIGHTS RESERVED
DOI: 10.1016/S1387-2656(05)11001-1
2

Introduction

These days, deciphering the rules that govern complex biological systems is one
of the greatest challenges to scientists. Complex systems in biology include
transcription regulatory networks in living cells, the patterns of cell division
and cell death that lead to the development of multi-cellular organisms, the
coordinated interplay between the various developmental stages of the vast
number of cells, the concerted response to innumerable environmental stimuli
and stress conditions, and, most notable, the coordinated communication among
the billions of cells of the human brain.
It has been argued that a better understanding of the fundamental biological
phenomena underlying disease processes would greatly facilitate the develop-
ment of novel therapeutics. Moreover, quantitative models on the cellular, tissue
and organ level would enable the prediction of the effects of therapeutic
interference, based on the drug’s chemical structure, the patient’s genetic makeup
and external environmental factors. It is expected that quantitative models of
diseases will revolutionise the drug discovery process, as such an approach would
enable a directed approach to the discovery and optimisation of bioactive
molecules as potential therapeutic agents. Likewise, potential drug safety issues
could be identified at a very early development stage by simulating the effect of
the drug candidate on the human organism. Using such a virtual research
environment, scientists could simulate large-scale experiments in silico that could
take months or years to do in the laboratory or in clinical research.
Obtaining a comprehensive quantitative understanding of the tremendous
complexity of life looks – at a first glance – like a hopeless enterprise. However,
over the last decade, a number of high-throughput experimental technologies
have been developed that enable a highly automated, comprehensive monitor-
ing of the various molecular levels of a cell (often referred to as ‘‘-omics
technologies’’). These global, large-scale experimental techniques are fundamen-
tally different from the traditional low-throughput technologies. A central
accomplishment was the development and implementation of high-throughput
DNA sequencing technologies that have enabled the automated sequencing
of whole genomes. Equally important, DNA chips and microarray technologies
are now available that facilitate the parallel quantification of all messenger
transcripts in a cell, the so called transcriptome. Analogously, proteomics and
metabolic profiling technologies have been developed to identify and quantify all
proteins and metabolites in a cell, respectively. Advanced technologies have
become available to investigate higher-order protein–protein interactions. In
parallel, gene inactivation techniques enable a systematic screening of disease-
and treatment-relevant phenotypes. Thus, for the first time in history, a sys-
tematic investigation of biological systems, ranging from a single cell to complex
organisms has become experimentally amenable.
Since the generation of experimental data has been greatly facilitated,
the current bottleneck lies in the biological interpretation of the data.
 3

Straight-forward data interpretation in a specific disease-context is hampered by

a variety of factors. In this review, we will discuss the most notable hurdles,
which include the lack of data integration systems, the lack of tools to reconstruct
the only partially known disease-pathways, and the lack of predictive models to
simulate the dynamics of biological systems for disease and treatment conditions.
As all these aspects are heavily dependent on information technologies, we
will present and discuss some of the most recent advances in algorithm, database
and simulation tool development. It will be demonstrated that computational
biology techniques have already a major impact on today’s pharmaceutical
research, including the development of disease hypotheses, the systematic identi-
fication of novel drug targets, and the in silico evaluation of new drug candidates.

Drug discovery: Bridging chemistry and biology

The industrialised pharmaceutical discovery process

Most drugs are small molecules that exert their action through modulation of
protein activity. New drug classes now arise mostly through industrialised
discovery programs that begin with the identification of a biomolecular target of
potential therapeutic value through biological studies, revolving around a so
called ‘‘disease hypothesis’’. With sufficient information about the proteins that
are critical for a functional pathway, these proteins are prioritised as potential
drug targets. Using enzymatic- or reporter-assay systems, chemical libraries are
tested for small-molecules that bind to the disease-involved proteins. The latter
process has been highly automated and is referred to as ‘‘high-throughput
screening’’ (HTS) [8]. For this, molecular libraries representing a large chemical
diversity are screened for active compounds (so called ‘‘hits’’). The compounds
identified in the screening campaigns are typically non-optimised structures, that
are weak binders and likely have a non-optimised pharmacokinetic profile. These
‘‘hits’’ now have to be transformed into ‘‘leads’’. Leads are defined as structures
that have been derived from an early hit, and although still not fully optimised,
show already some appropriate characteristics of a precursor of a drug. A good
lead often has some proof-of-concept activity, but will likely not have been fully
optimised for pharmacokinetic properties and off-target activities. Ultimately,
the leads have to be optimised in a cycle that features design, synthesis and
assaying of numerous analogs.
There are many hurdles and subtleties along this way, and many promising
drug candidates fail at one or the other point along the development process.
This makes the development of a new drug a tedious, complex and very expensive
enterprise. Over the last few years, the average time for development has
dramatically increased now to 75 months [80]. But even of those compounds that
reach the later stages of development, more than 80% will not be approved
for sale [80]. Consequently, the number of new drug approvals has remained
relatively constant, with only approximately 30 new molecular entities approved
4

per year, despite a doubling of the spending for pharmaceutical research and
development over the last decade [3]. The whole process of developing, testing
and obtaining approval for a new medicine costs now an average of approx-
imately $900 million for each new drug that makes it to the market, taking into
account the candidate compounds that fail along the way [80].

Major knowledge gaps in biology: Bottleneck for target discovery

At first sight drug discovery appears to be an entirely chemistry-driven problem,

as the ultimate products of the pharmaceutical industry are typically chemical
entities (i.e., drugs). However, one of the major reasons for the low number of
successful drug discovery projects is their high attrition rate during development,
particularly from issues related to failures of on-target biological hypotheses and
on- and off-target safety concerns. As long as the underlying physiology of drug
targets is not fully understood, there will be many instances of failure from
incorrect biological hypotheses. To address these problems that often occur in the
later stages of a drug development programme or even during clinical tests, there
has been a recent emphasis on optimizing the earlier stages in the drug discovery
process, with a key problem identified as the choice of appropriate targets [3].
The currently marketed drugs interact with only approximately 400 genes
or gene products, with an estimated several thousands of genes being impor-
tant genes for disease predisposition, onset and progression [9]. Interestingly,
the number of functionally uncharacterised human genes is still very high, with
estimates ranging from 30–50% genes coding for a gene product of unknown
function [37,56]. This is not only true for the human genome, but also for well-
investigated, biomedically relevant model organisms, such as mouse, rat,
Drosophila melanogaster, or Caenorhabditis elegans. Even for unicellular euka-
ryotes (such as yeast) or the bacterial genomes there are typically more than 30%
unknown genes.
Unfortunately, there is also little known about the pathways and functional
cellular systems that are of central importance for pharmaceutical applications.
One example of an important target class whose functional context is largely not
known are G-protein coupled receptors (GPCRs). GPCRs are a family of
receptors that mediate most of the cell–cell communication in humans. The
broad variety of extracellular activators and ligands, such as hormones, ions,
neurotransmitters, that signal through GPCRs underscore the physiological
importance of this receptor class. GPCRs are involved in many important bio-
logical functions, such as photo- and chemoreception, neurotransmission,
regulation of endocrine secretion, blood pressure control, embryogenesis, angio-
genesis or tissue regeneration, just to name a few [11,12]. The ability of molecules
to selectively interact with GPCRs, together with the selective nature of GPCR
gene expression has provided the basis for the pharmaceutical relevance of this
receptor family. It has been estimated that 50% of all modern drugs and about
one-quarter of the top 200 best selling drugs modulate GPCR activity [13].
 5

Overall, about 650 GPCR-coding genes could be identified by sequencing the

complete human genome. Roughly 190 of those are categorised as ‘‘known’’
GPCRs, as they are known to be activated by some 70 identified ligands [12].
Interestingly, the currently available GPCR-based drugs target only about 30%
of the known GPCRs [12]. One major outcome of the human sequencing project
was that there may be further receptor subtypes of known and entirely novel
GPCRs that could be utilised as drug targets. A new receptor is called ‘‘orphan
receptor’’, if its endogenous ligands remain to be discovered. These orphan
GPCRs are considered to be of great promise for the development of novel
therapeutics, which explains the increasing attention from the pharmaceutical
industry. As most drug screening technologies rely on the identification of small-
molecules that interfere with ligand binding to GPCRs [14], a necessary pre-
requisite for a discovery programme is the identification and characterisation of
receptor–ligand interactions, and to get a better understanding which down-
stream signalling cascades are activated. This is critical information to come to
well-founded disease hypotheses, preventing expensive failures during later-stage
discovery and development programmes [3].
Another example of an important target class among signalling molecules are
kinases [10], an enzyme class that shares a catalytic domain conserved in
sequence and structure. The sequencing of the human genome showed that there
are approximately 518 genes encoding kinases [15]. Protein kinases mediate most
of the signal transduction in eukaryotic cells, controlling many cellular processes
such as cell cycle progression, apoptosis, cell differentiation, transcription and
metabolism. Mutations and dysregulation of protein kinases are known to play
causal roles in human diseases, affording the chance of developing agonists and
antagonists of these enzymes for use in disease therapy. Therapeutics based on
kinase inhibitors already play a central role in various medical indications, most
notably for cancer therapies [16,19].
The majority of the currently known kinases are little characterised in terms
of the cellular context they act. Most textbook pathway maps describe the
kinase-mediated signalling cascades as linear chains from the cell membrane
reaching down through the cell into the nucleus, ignoring signalling redundancy
and pathway crosstalk. In reality the connectivity of signalling pathways
has major implications on defining a disease hypothesis and selecting a suitable
target. For instance, there is much dispute about as to where to optimally
interfere in an oncogenic signalling pathway to maximise the anticancer effect
and minimise the toxicity to normal tissues [4]. Interfering with the central
machinery for cell-cycle control, upon which multiple signalling cascades
converge, would probably have a powerful anti-cancer effect, but most likely also
have strong effects on normal cells. Based on this, it would follow that blocking a
membrane receptor (e.g., a receptor tyrosine kinase) would be preferable,
however, one could argue that too many disease-unrelated downstream signals
would be affected. Hence, interference at an intermediate level could be seen as
optimal. In reality, we still do not know enough about the precise arrangement of
6

the oncogenic wiring of individual cancers to make a rational selection of a target

molecule [4].
The GPCR and the kinase examples show that characterizing the functional
and potential pathophysiological roles of drug targets is a necessary prerequiste
for a rational target prioritisation strategy and successful follow-up drug
discovery projects. As the majority of the tremendous complexity of the human
signal transduction systems is still unknown, it has been proposed that more
molecular data in conjunction with new in silico interpretation strategies could
help in efficiently prioritizing the most promising target candidates.

The ‘‘-omics’’ technologies

A whole panel of newly developed technologies are about to revolutionise the

experimental data acquisition process in biological sciences. These technologies
enable the comprehensive characterisation of the genome, transcriptome, pro-
teome, and metabolome (see Fig. 1). What is common to all these technologies

shotgun
sequencing
Genome

Transcription

microarrays,
oligonucleotide chips
Transcriptome

Translation

2-D gels, ICAT,

Proteome
mass spectrometry

Reactions

mass spectrometry
Metabolome
(GC/MS)

Interactions

yeast-2-hybrid
Interactome
screens, TAP

Integration

Gene inactivation,
Phenome
Knock-outs, RNAi,…

Fig. 1. The basic molecular levels of a cell, together with the biological processes
communicating between these levels. The boxes to the left indicate representatives of newly
developed ‘‘-omics’’ technologies (gray) that can quantitatively monitor the respective
biological levels in a highly automated fashion [155].
 7

is their high degree of automation. Here, a brief overview on the most widely-
used large-scale experimental techniques in pharmaceutical research will be
given.

Genome sequencing technologies

Not even a decade after the sequencing of the first bacterial genome, the
determination of complete eukaryotic genomes has become a straightforward
procedure. Most notably, the complete human genome has been sequenced
[37,56], together with hundreds of other pharmaceutically relevant organisms,
including important model organisms like rat, mouse or pufferfish. Additionally,
the genomes of virtually all major bacterial, fungal and viral pathogens have also
been published in the meantime. While currently the focus of many sequencing
laboratories is to produce genome drafts of phylogenetically very different
species, the future will likely bring the sequencing of various strains or individ-
uals of the same species, aiming at the investigation of the genetic variability and
its relationship to disease predisposition and treatment susceptibility. In this
context, microarrays for detecting and characterising single nucleotide poly-
morphisms (SNPs) will play an increasingly important role in genotyping [2,7].

Transcriptomics technologies
DNA microarray technologies permit a systematic approach to obtaining
quantitative information on the transcriptional activities of all genes in a cell.
Investigation of the transcriptome reveals how global gene expression is
remodelled during changes in cell growth, physiology or environment. Currently,
two major types of DNA chip technologies are being used that can be broadly
termed as ‘‘one-channel’’ and ‘‘two-channel’’ technologies [41]. While one-
channel technologies were developed to measure the absolute concentration of
mRNA transcripts, two-channel technologies measure the relative abundance
between sample and control specimen. Several different DNA microarray
technologies are currently in use: some DNA chips make use of partial or
complete cDNA, or, alternatively, utilise chemically synthesised oligonucleotides
that represent the genes whose expression activity is to be monitored [42].
Alternatively, pre-fabricated DNA molecules can be printed in arrays on glass
slides or nylon membranes. Typical spot sises on a microarray have shrunk
significantly, enabling chip designs that cover complete gene sets of complex
organisms. For instance, the currently commercially available microarrays can
now measure up to 47,000 different trancripts in parallel [1].

Proteomics technologies
Since the proteome of a cell is not a simple reflection of its transcriptome, direct
protein-based technologies are needed. Changes in protein isoforms due to post-
translational modifications, such as phosphorylation induced by cell signalling
events are known to be relevant for understanding disease processes. Advances
in proteomic technologies are now improving the quantification of membrane
8

proteins and signalling complexes with increased speed and molecular detail
[45]. In contrast to nucleic-acids-based technologies, the chemical and structural
diversity of proteins pose major obstacles for the development of technologies
designed to measure the abundances of all proteins in a cell simultaneously.
Besides protein chips and mass spectrometry technologies [43], the most
commonly used technology is 2D-gel electrophoresis. In the electrophoretic
approach, proteins are separated according to their molecular weight and
isoelectric point. In contrast to DNA chips, where a measured signal can be
readily assigned to a specific gene, the proteins represented by the spots on
2D-gels first need to be detected and the proteins identified. Technology-inherent
limitations must be resolved before proteomics in general, and 2D-gel electro-
phoresis in particular, becomes a widely used, industrial-scale platform [20].

Metabolomics technologies
The quantitative monitoring of all metabolites and metabolic flux patterns in a
cell is increasingly attracting attention [44]. Important biological effects can only
be explained on the metabolism level, such as e.g., gene regulation mechanisms
that are controlled primarily by cellular metabolite concentrations. Although
the detailed knowledge of low-molecular weight metabolites in a cell is of utmost
importance for various biomedical applications, metabolic profiling technol-
ogies are still in their infancy. Currently, it appears that the most mature
metabolomics technology is a gas chromatographic separation of compounds
with a subsequent mass spectrometric identification (GC/MS) [46,47]. Small
molecules are ionised and separated based on their mass-to-charge ratio.
Measuring the molecule’s masses of distinct fragments formed during ionisation
leads to characteristic chromatographic peaks.

Interactomics technologies
Over the last few years, a number of different technologies have been established
to experimentally determine protein–protein interactions. The most widely used
technology are clearly the yeast-two-hybrid systems (Y2H), based on a fusion
of the protein of interest to a transcription factor DNA-binding domain that
can bind to the promoter of a reporter gene [27,28]. Alternatively, one may
use in vitro biochemical methods, such as fusing the protein to a tag, such
as glutathione S-transferase (GST), which enables to ‘‘pull down’’ possibly
interacting proteins [22]. A commonly used in vivo biochemical method is
co-immunoprecipitation. Starting from a specific antibody, proteins are
immunoprecipitated, which enables the identification of interaction partners by
western blotting or mass spectrometry. Alternatively, protein arrays with
immobilised proteins in array format can directly probe for interacting proteins
[24]. Y2H assays have been shown to be able to detect also transient interactions
(i.e., typically weak interactions for only a short time, such as a kinase–substrate
interactions in a signalling reaction), whereas the pull-down methods identify
mostly the stable protein compexes [25]. The first genome-wide screen for
 9

interacting proteins for yeast has been published in the year 2000 [29,30].
Currently, complete interaction maps have become available for a few uni- and
multi-cellular organisms (e.g., D. melongaster) [31,33,34]. Large-scale protein–
protein interaction screens can be used to systematically reconstruct signalling
pathways (see Fig. 2). Although the discussion is still ongoing as to what extent
the results produced by the different technologies are comparable (see e.g.,
[25] and references therein), it has been recognised that systematically measured
protein–protein interactions provide valuable insights into higher-order biolo-
gical phenomena [26].

Phenomics technologies
Gene inactivation technologies play an increasingly important role in experi-
mentally screening for pharmaceutically relevant phenotypes. A gene can be
either mutated or inactivated by direct gene disruption or RNA interference
(RNAi). Some technologies can be applied on a whole-genome scale, providing
comprehensive information about specific phenotype categories [1,35,36,51].
A major difficulty of phenotype screens is that it seems that many coding
regions code for products whose loss does not result in an obvious macroscopic
phenotype. Most likely, this is due to the inherent functional redundancy of the
different gene products. For instance, less than 20% of all yeast gene products
are essential [49]. As long as yeast cells are grown on rich media, it is difficult to
observe any phenotype that could provide a hint towards the function of the
missing gene product, indicating that it is difficult to find relevant developmental
and environmental conditions that would reveal interpretable phenotypes.

Data generation vs. data interpretation

It is instructive to discuss how the availability of high-throughput technologies

has already changed our way to do biological research. This shall be exemplified
by first focussing on high-throughput DNA sequencing technologies. Note that
similar considerations are true also for the other technologies, in particular
transcriptomics, proteomics and metabolomics.
In retrospect, there was very little DNA and protein sequence data available
from the 1950s on. The invention of automated DNA-sequencing technologies in
1975 changed the situation dramatically [59]. In the majority of cases, protein
sequences were now no longer directly determined, but rather derived as so called
‘‘conceptual translations’’ from a DNA template. However, DNA sequencing
was still a tedious task, and in most cases a study started out with a so called
‘‘functional cloning’’ or a ‘‘positional cloning’’ project and the DNA sequence
was determined afterwards. As in the early days of sequencing there was so
little DNA data, a sequence similarity search resulted in most cases not in the
identification of a similar sequence. Over the last decades, approximately
ten thousand genes were functionally characterised by starting from a functional
 10
Fig. 2. Integrating biological data produced by different high-throughput experimental technologies. In this example, signalling pathways have
been reconstructed to investigate their activity patterns for different cancer stages. (a) The complex signalling network of the human RAS–RAF
pathway, reconstructed from experimental protein–protein interaction data. Proteins (yellow circles) are connected by lines representing
experimentally determined physical protein interactions. The different colors indicate various technologies used in the experiments (e.g., Y2H,
TAP or co-immunoprecipitation). (b) Automatically structured EGF signalling pathway with integrated and overlaid microarray data. The lines
represent experimental protein interaction data, while the coloring of four horizontally arranged squares next to each of the proteins represent
the corresponding transcript fold changes measured across four microarray experiments, each representing a different patient (red=upregulated,
green=downregulated, black=non-responsive); PhylosopherTM, Genedata, Basel.
 11

5 sequences
Sequences / publications [Mio]

4
Widening gap
3

publications
1

1975 1980 1985 1990 1995 2000

Year

Fig. 3. Cumulative increase of published articles in molecular biology and genetics,

representing mostly classical experimental studies, compared to the increase of DNA sequence
records in GenBank. Publications are shown as a blue dashed line, while the green solid line
represents the number of gene sequences. The numbers on the y-axis can be read as ‘‘genes’’,
although no attempt has been made to eliminate redundancies in either GenBank or the
literature databases (modified after [58]). Note the widening gap between data produced by high-
throughput experimental technologies and traditional studies. A similar dynamics can be observed
also for transcriptomics data, proteomics data, metabolomics data and interactomics data.

or positional cloning project, and a subsequent sequencing of the relevant

DNA [58].
In the early 1990s, there was more and more sequence data available, mostly
due to the launch of the Human Genome Sequencing project, and the frequency
of database search hits increased dramatically. The critical turning point came
in 1995, when the number of genes in public databases began to exceed the
number of papers in the scientific literature (see Fig. 3). Currently, the number of
completely sequenced genomes grows exponentially, with a doubling time of less
than 1.5 years (see Fig. 4). As a consequence, the gap between low-throughput
traditional experimental approaches (e.g., cloning) and the exponentially
growing sequence databases is expected to widen dramatically.
The standard annotation procedures via homology (i.e., sequence similarity)
used to work fine before the early 1990s, as whenever a sequence similarity search
against a DNA database turned out to be successful, the identified similar
sequence typically was associated to a wealth of other experimental findings
indicating the function of the gene or protein (e.g., derived from a functional
cloning project). After the mid-1990s, the situation changed, as there were many
more sequences available compared to the relatively slowly increasing number of
12

H. sapiens
200

100 H. influenzae

50 S. cerevisae
Complete genomes

20

10

1
1996 1998 2000 2002 2004

Year
Fig. 4. The number of publicly available complete genome sequences as a function of time.
Arrows indicate the publication years of H. influenzae as the first completely sequenced
genome, S. cerevisiae as the first completely sequenced eukaryotic genome, and the human
genome. Note that the vertical axis represents a logarithmic scale. The straight line is a fit of an
exponential growth function with a doubling time of less than 1.5 years (modified after [83]).

traditional experimental studies. The exponentially accumulating sequence data

led to a decrease in the average informativeness of the annotation-by-homology
[58], sometimes referred to as the homology barrier.
The exponential increase in public and proprietary databases of microarray,
proteomics and metabolomics data obeys a similar exponential dynamics as the
DNA sequence databases. Thus, an analogous discrepancy between the dynam-
ics of data generation and the efficiency of data analysis and interpretation
exists. The heterogeneity, complexity and volumes of the biological data
that is generated these days prevents a straight-forward exploitation of the
data in drug discovery projects. Thus, computer-based systems are expected
to play an increasingly important role for storing, structuring and analyzing
research data. While over the last few years many isolated tools for specialised
tasks have been developed, there are only a few attempts to develop inte-
grated computational systems that address the diverse requirements related to
data-driven drug discovery applications (see Fig. 5). In the foreseeable future the
 Input: Public & Data upload Relational Scientific Result and Output: pathway
proprietary and integration database analysis and project reconstructions
data interface system visualisation management and simulations
Fig. 5. Schematic overview of the central requirements for a computer-based system supporting a data-driven drug discovery process.
Experimental data has to be quality assessed, before it is uploaded into a central database. Within the database, the experimental data need to be
automatically connected with public and corporate information about disease-relevant genes, transcripts, proteins, functional information,
pathways, as well as tissue, animal and clinical data. Integrated, scientifically validated tools support researchers to mine the data and to
elucidate functional relationships and biological networks. Intuitive and interactive visualisers are needed for biologically interpreting the data

13
to come to experimentally testable hypotheses. In addition, the system must provide a mechanisms to share and distribute findings, conclusions
and evidences on disease-hypotheses, target proposals, screening assays and bioactive compounds. Integrated result-sharing facilitates the
coordination of multiple target, assay and compound projects across different therapeutic areas and research sites.
14

availability of integrated enterprise systems will be crtitical to take full advant-

age of the rapidly growing biological data and the recent methodological
progress in computational biology.

Inventorizing biology: An encyclopaedia of life

Complete genome sequences

The release in 1995 of the first complete genome sequence of Haemophilus

influenzae marked the advent of a new age in biology [38]. For the first time ever,
the complete genetic information of a free living organism was revealed.
Complete genome sequences are indispensable for cataloguing all broadly
conserved housekeeping genes, as well as for understanding the genetic basis for
the different species and phylogenetic lineages. Furthermore, complete genome
sequences are necessary to ascertain that a certain protein implicated in a specific
function is not encoded in a given genome. These are the major advantages of
having a complete genome sequence available, explaining the public excitement
when the first draft of the human genome was published in 2001 [37,56].
Nowadays, newly sequenced genomes are being published almost on a daily
basis. A survey of the publicly available genomes gives some insight into the
dynamics of the ongoing sequencing projects, and provide also a perspective on
what can be expected in the next few years. Figure 4 shows that the number of
genomes is exponentially increasing, with a doubling time of approximately 16
months [83]. This data suggests, that within a few years thousands of genomes
will become available.
Although being extremely useful, the sequencing technology itself just
generates long, cryptic nucleotide sequences, consisting of a few millions to
several billons of As, Ts, Gs and Cs. This sequence raw data will only be useful if
it can be biologically interpreted, i.e., that the encoded gene products can be
identified, and their molecular and cellular function be characterised. This
means, we have to understand what the consequence of a certain sequence or a
mutation variant thereof means for the physiology of an organism, possible
disease states and treatment conditions.

What is a gene?

Most functions in a cell are taken over by proteins, encoded by ‘‘genes’’.

Historically, a ‘‘gene’’ was defined as an abstract concept to explain the
hereditary basis of traits [48]. The phenotypic traits were ascribed to hereditary
factors, although it was not clear what the physical basis of those ‘‘genes’’
should be. The advent of recombinant DNA and gene cloning techniques made
it possible to associate a ‘‘gene’’ to a specific piece of DNA, and its related
gene product, being either a protein or a nucleic acid. A modern, molecular
definition of a gene is ‘‘a complete chromosomal segment responsible for making
 15

a functional product’’ [49]. According to this definition, there are different

criteria that are commonly used for finding genes in genomic sequences using
computer-based techniques.

Open reading frames

One way to readily identify genes in genomic sequences is to determine all open
reading frames (ORFs), i.e., DNA sequence regions that are operationally
defined as long sequence segments bound by a start and stop signal triplet. This
approach to gene finding works well for prokaryotes, and organisms with no or
little splicing. However, in organisms that make extensive use of splicing for
modifying their primary transcripts, ORF identification can be very difficult.
Another complication are small genes which are frequently overlooked by clas-
sical ORF finding.

Sequence features
Genes can also be identified by investigating characteristic signatures in the
DNA sequences. One example is the typical codon bias indicating functional
genes. Although the codon bias is often very weak, this method helps in many
cases to discriminate false-positive ORF predictions from real genes. Also,
splicing sites are frequently indicated by specific sequence patterns that can be
used to predict genes [50]. However, when relying entirely on DNA sequence
features, the best algorithms predict typically fewer than 50% of the true exons,
and less than 20% of all genes [52].

Sequence conservation
As more and more genomes are being sequenced, the comparison of the genomic
sequences across different species can provide valuable insights into which parts
of a genome are conserved and therefore likely code for functional genes. The
species that are used for finding genes have to be carefully chosen, as they have to
be separated by appropriate evolutionary distances. Obviously, species-specific
genes cannot be identified this way.

Evidence of transcription
By using additional experimental evidence coming from independently
sequenced mRNA or protein sequences, transcribed regions in a genome can
be pinpointed. Particularly important in this context are sequenced mRNA
messengers and EST sequences. By mapping those sequences onto completely
sequenced genomes, alternative splicing variants can be identified. Alternatively,
microarrays harbouring sequences of entire genomes or chromosomes can be
used to identify stably expressed genes [53,54].
There are a variety of different gene finding programmes available, some of
them using ab initio approaches, some using comparative approaches, while
others combine different criteria. Still, even for well-investigated genomes like
the human genome, there is little agreement on the exact number of genes.
16

For instance, the estimates for the human genome still range from 24,500 to
45,000 genes, depending on the specific gene finding programme used [156].
There is some convergence in the gene counts observable, with a predicted
100,000 human genes only a few years ago.
A discussion of all possible issues related to gene prediction is beyond the
scope of this review. Typical difficulties in identifying genes arise through a
number of effects. For instance, gene overlaps occur frequently (i.e., exons of
genes that are located in the intron of another gene, or overlapping protein-
coding and RNA-coding genes, see e.g., [57]). Another hurdle in identifying
functional genes are the abundant pseudogenes, i.e., sequences that are similar
to normal genes, but which usually contain a frameshift or stop codon in the
middle of the coding sequences. It is estimated that among the currently
‘‘known’’ human genes there are several thousands of pseudogenes. Lastly,
extensive alternative splicing typically encountered in higher eukaryotes can
obviously complicate matters.

Homology-based functional annotations

Proteins are the main catalysts, structural elements, signalling messengers and
molecular machines of biological tissues. Thus, after having sequenced a
genome, the encoded proteins need to be functionally annotated to make the
genome content accessible to biological interpretation.
In a first step, the protein coding genes are typically translated into ‘‘virtual’’
protein sequences using a genetic code translation table (‘‘conceptual trans-
lations’’). After having produced this ‘‘theoretical proteome’’ (i.e., the complete
set of all proteins that an organism can produce), there are various ways how
functional annotation information can be assigned to the newly predicted
proteins.
The traditional way to learn about the molecular function of a newly identified
gene is to use its sequence and compare it against a repository of all other
sequences known so far, in the hope to find a match to a sequence that can be
associated with some experimental evidence of its function. This ‘‘homology-
based’’ method acts on the assumption that all primary knowledge comes from
some biochemical, structural or genetic experiment of an individual protein. The
reasoning behind this straight-forward annotation approach is that once some
function is assigned to a protein sequence, one can easily identify similar
sequences whose products are likely to share the same function. In technical
terms, there is a wide range of algorithms that can be used for efficient sequence
similarity searches. By far the most popular program is BLAST, a widely-used
program to screen large sequence databases for similar sequences [157].
A more general approach to homology-based annotation of newly sequenced
genomes is based on using sequence motif ‘‘signatures’’ instead of the actual
sequences. Motif models of functional protein domains can be extremely helpful
for eludidating the function of uncharacterised proteins. Motif-model based
 17

methods have the advantage of being more sensitive as only the conserved and
thus functionally more relevant residues enter the analysis. A number of mostly
academic consortia are compiling functional protein domains that are
represented as mathematical representations of a sequence consensus (e.g., as a
so called ‘‘Hidden Markov Model’’). One of the most popular database of
functional protein motifs is PFAM [158]. Such motif libraries can be readily
used to predict the molecular function of newly identified proteins.
Homology-based methods are also sometimes used in a more general context
for annotating cellular functions. This means that proteins that share some
sequence similarity are predicted to belong to the same pathway, a similar
cellular process, or that the proteins are localised in the same cellular
compartment. There are many examples that show that sequence similarity
indicates primarily the molecular and not the cellular function. For instance,
many kinases can be very similar in sequence, but are involved in very different
signalling pathways. Analogously, phenotype information associated with a
sequence (e.g., ‘‘essential for cell division’’) is often transferred to proteins with
a similar sequence from another organism. Annotations of this type have to be
used with caution.

Transcription body atlas

Gene transcription in its full complexity is an only poorly understood process.

Transcription of genes is tightly controlled in all organisms during develop-
ment and within different cell types, and can respond to changes in environ-
mental stimuli. Moreover, most genes of higher eukaryotes undergo complex
splicing. This means, that the primary transcripts are further processed and
modified before the mRNA is translated into a polypeptide. Splicing processes
are often tissue- or development-stage specific. For some human genes dozens of
alternative splicing variants are known. About half of the human genes have
spliced isoforms, and this is likely to be an underestimate, as not all variants
have been identified so far [55,56]. To investigate the variety of splicing pro-
cesses, mRNA sequences have to be clustered and mapped onto the complete
human genome to define a non-redundant set of all messenger RNAs.
In these days, microarrays are increasingly used to get a comprehensive
picture of the anatomical distribution of transcription under ‘‘normal’’ physio-
logical conditions of humans and relevant model organisms [140], and changes
thereof due to disease conditions or treatment effects (see e.g., [160,159]). Gene
expression within each of the many different human cell types is characterised by
a distinct gene expression pattern. These cell-type specific expression patterns are
of course modulated in time, partly due to long-term processes such as develop-
mental signals, due to short-time processes such as oscillatory cell division
processes, and lastly because of (typically non-regular) external influences such
as stress conditions.
18

A complete catalogue of the temporal and spatial expression patterns in

humans will likely help to reveal novel drug targets, in particular when combined
with information about specific protein classes considered as ‘‘druggable’’, such
as e.g., GPCRs (see Fig. 6). The integration of large-scale expression data with
sequence homology-based data is already used to obtain a more complete
understanding of gene function. For example, two orphan GPCRs, GP31 and
GPR9, both show enriched expression in the pancreas, suggesting a role of
these proteins in digestion or hormone secretion. The tissue-specific expression
of target proteins is an important selection criteria for therapeutic targets, as
the primary effect of modulating their function will likely be restricted to the
target tissue [140].
Comprehensive transcriptome data is a basis for characterizing functional
relationships between proteins and enzyme–substrate pairs. For instance, the
identification of the cognate ligands for orphan GPCRs is an important
prerequisite to set up screening assays for lead identification. The anatomical
distribution of expression provides valuable information related to the potential
physiological function and therapeutic utility of GPCRs [12]. For instance,
large-scale expression data was used to show that GPRK2L, a GPCR kinase,
is expressed almost exclusively in testis. The same microarray data revealed
that there are only fifteen GPCRs that are detectably expressed in testis. Thus,
these GPCRs represent the most likely substrate candidates for GPRK2L [140].

Genome–genome comparisons

The availability of a whole range of phylogenetically diverse complete genomes

enables the delineation of comprehensive cross-species gene and protein
families. Such family information is critical for comparing the set of encoded
molecular functions by a given genome, and to put this ‘‘parts list’’ in the context
of other completely sequenced genomes. The relationships between the genes
of different organisms are represented as a system of homologous families,
including so called orthologs and paralogs. Orthologs are genes descending
from a common ancestral gene, while paralogs typically originate due to gene
duplication events in the course of evolution. Normally, orthologs retain
their function, whereas paralogs take over new functions, typically related to the
original one. On a higher level, pathways may also be compared across dif-
ferent species to pinpoint species-specific differences in signal transduction and
metabolism.
From a drug discovery standpoint, comparing organisms on a pathway level
enables the evaluation of organisms as models for human diseases. In some
cases, even phylogenetically only distantly related organisms are considered as
suitable models. For instance, D. melanogaster has been discussed as a model for
some aspects of Alzheimer’s disease [161]. It is important to have a thorough
understanding of the similarities and differences between humans and the model
 kidney cerebellum, amygdala, cortex, DO-H2, GA10, Whole
caudate nucleus, thalamus, HL60, K422, blood
corpus callosum, fetal brain, WSU
spinal cord

Transcript concentration [arb. units]

GPCR mRNA profile similarity

Tissue samples
Fig. 6. A heat map of the differential expression activity of human G-protein coupled receptor encoding genes across a broad array of human
tissues and cell types. GPCR genes were systematically identified by screening all human gene products for characteristic GPCR sequence
signatures (here represented by significant PFAM matches of the PF00001, PF00002 and PF00003 motifs, see e.g., [158]). The corresponding
hybridization activities measured in microarray experiments are color-coded, representing the respective transcript concentration, with the GPCR-genes
represented by rows, and the different tissues and cell types by columns. The clearly visible block structure indicates tissue-specific expression patterns of

19
many of the GPCR genes. The selective nature of GPCR gene expression is one important reason for the pharmaceutical relevance of this receptor
family. For clarity, only a subset of all GPCR-related probes is shown (data from [140], analyzed in PhylosopherTM, Genedata, Basel).
20

organisms before starting large-scale animal studies, but also for interpreting the
relevance of experimental findings for humans.
Shortly after the availability of a handful of completely sequenced genomes
the first systematic framework for classifying and categorising protein families
across multiple genomes was published [64]. The originally proposed approach is
based on so called ‘‘clusters of orthologous groups’’ (‘‘COGs’’). The COG
protein families are constructed using many-to-many relationships that are
derived by an all-against-all comparison of all protein coding genes from
different species. Importantly, the proposed algorithm works also for identifying
orthologous relationships across large phylogenetic distances. The result is a set
of protein families, with each family typically representing one (or few) distinct
molecular functions. Thus, the full set of COGs can be considered as a
representation of all elementary building blocks of life.
Applying the COG algorithm to the seven complete genomes that were
available in 1997 (including gram-negative and gram-positive Eubacteria,
Cyanobacteria, Archaea, and yeast) resulted in slightly more than 700 COGs.
The expansion of this system by incorporating additional 27 genomes, including
the multicellular C. elegans and D. melanogaster, led to an increase of the
number of protein families to approximately 2,800 COGs [65]. With the inclu-
sion of more genomes, the discovery of additional protein families will gradually
level off, as the majority of the genes encoded in newly added genomes will fit
into already existing COGs [64,65].

Reconstructing pathways I: Predicting cellular functions from contextual

genome information

Beyond homology

Inherent in the growing collection of genome sequences is the knowledge about

functional linkages between proteins, reflecting the cellular function of the
proteins. Structuring the ‘‘gene content’’ of complete genomes across a wide
range of different species does not only allow the direct comparison of the
building blocks of a species, but can also reveal functional relationships between
genes and proteins.
As discussed above, the traditional functional annotation methods are mostly
relying on similarities between amino acid sequences. In many cases, however,
one is faced with a highly conserved set of proteins, for which no experimental
evidence exist that could indicate the protein’s function. Recently, a number of
computational methods have been developed that go beyond the classical
homology-based annotation approaches, as they can supply functional informa-
tion on fully uncharacterised proteins. A major outcome of these methodologies
is that they provide ‘‘functional links’’ between different proteins. In many cases
this is valuable information to put uncharacterised proteins into a functional
context, such as a metabolic pathway, a signalling cascade or a protein complex.
 21

Even if the proteins ‘‘linked together’’ are all of unknown function, these
methods provide valuable hypotheses for future experiments [40,105].

Phylogenetic profiling analysis

A so called ‘‘phylogenetic profile’’ characterises the absence or presence of a

conserved protein across a set of organisms whose genomes have been
sequenced. The central assumption of the phylogenetic profiling approach is that
if two proteins have the same (or at least a very similar) phylogenetic profile
across a large number of genomes, then there is a predicted ‘‘functional link’’
between the two proteins. This statement is equivalent to the hypothesis that a
set of genes coding the building blocks of a functional system (e.g., enzymes
catalyzing adjacent steps in a metabolic pathway, or proteins that make up a
protein complex) are inherited together, to obtain all the required proteins for a
fully functional system. This hypothesis implies a correlation of the protein’s
underlying phylogenetic profiles. A typical phylogenetic profile correlation
pattern is shown in Fig. 7. The selected patterns are very similar, as they all
represent proteins required for the bacterial flagellum apparatus, a specific
cellular machinery including structural proteins, biosynthetic proteins, signal-
ling molecules and secretion proteins. It has to be pointed out, that this phylo-
genetic pattern correlation is fundamentally different to the correlation between
individual amino acid sequences that is commonly used to infer molecular
functions. Two distinct proteins with similar phylogenetic profiles are typically
very different in terms of their amino acid sequence similarity [40,61,105].
Currently, most phylogenetic profiling approaches are based on comparing
binary patterns, with each presence or absence of a given protein represented by
a ‘‘1’’ or a ‘‘0’’, respectively. This explains why phylogenetic profiling approaches
are expected to become more accurate the more genomes are analysed in parallel:
If there are n fully sequenced genomes, there are 2n possible phylogenetic profiles
possible. Currently, there are more than 200 fully sequenced genomes available,
resulting in 2200
1060 profiles, an astronomic number significantly exceeding the
number of protein families (typically a few thousands, see e.g., [65]). The large
difference between these two numbers means that binary phylogenetic profiles
are an almost unique signature of a given functional system. Hence, any two
proteins with the same or a similar phylogenetic profile are likely to be engaged
in a common pathway or multimeric structural complex.
In technical terms, the ‘‘presence’’ or ‘‘absence’’ of a protein encoded in a
genome can be defined by a simple sequence similarity threshold, starting from a
reference genome [61]. This concept has been generalised to continuous profiles,
reflecting the evolutionary distances of the proteins contributing to a profile [62].
The usage of reference genomes obviously introduces a bias, as genes not
encoded in the (typically arbitrarily chosen) reference genomes cannot be
represented this way. An approach independent of any reference species is based
 22
FliR Flagellar type III secretorypathway protein
FliG Flagellar motor switch protein
FliF Flagellar type III secretorypathway lipoprotein
FliQ Flagellar type III secretorypathway protein
FliP Flagellar type III secretorypathway protein
FlhB Flagellar type III secretorypathway protein
Phylogenetic profiles

FlhA Flagellar type III secretorypathway protein

FliN Chemotaxis/type III secretorypathway protein
FliC Flagellin and hook-associated proteins
MotA Flagellar motor component
FlgK Flagellar hook-associated protein
FlgD Flagellar hook capping protein
FlgC Flagellar basal body rod protein
FliE Flagellar hook-basal body protein
FlgG Flagellar basal body and hook proteins
FlgB Flagellar basal body protein
FlgI Flagellar basal-body P-ring protein
FlgH Flagellar basal body L-ring protein
YbbJ Membrane-bound protein
FlgJ Flagellum-specific muramidase
Genomes
Fig. 7. Phylogenetic profile clusters suggesting functional relationships between non-homologous proteins. The phylogenetic profiles are
characterized by the presence (red) or absence (pink) of a protein across a large number of different complete genomes. The rows represent 20
proteins that show a similar phylogenetical distribution, indicating a related cellular function. Indeed, the available annotations (right; in bold
face the E. coli gene name) show that the proteins are all involved in the bacterial flagellar machinery. The set of proteins can be further
separated into two subgroups, one representing the secretion type III machinery necessary for functioning flagella, but also used in a different
context by other, non-flagellated bacteria (e.g., Chlamydia, Cpneu and Ctra, see box). Here, 19 phylogenetically very different genomes have
been compared: D. melanogaster (DmelN), S. cerevisiae (Scere), B. melitensis (Bmel), N. meningitidis (Nmen), E. coli (Ecoli), H. influenzae
(Hinf), H. pylori (Hpyl), B. subtilis (Bsub), S. pneumoniae (Spneu), M. tuberculosis (Mtub), Synechococcus (Syne), C. pneumoniae (Cpneu), C.
trachomatis (Ctra), B. burgdorferi (Bbur), T. maritima (Tmar), D. radiodurans (Drad), A. aeolicus (Aaeo), A. fulgidus (Aful), M. jannaschii
(Mjan); PhylosopherTM, Genedata, Basel.
 23

on a generalised COG-like family concept, with the phylogenetic profiles defined

by the ‘‘protein content’’ of the protein families (see e.g., [63]).
The functional links predicted by phylogenetic profiling are typically used to
predict the pathway context of proteins of unknown function. One application,
for instance, is the systematic identification of potentially involved new players in
metabolic pathways, based on their phylogenetic profile only (for an example,
see Fig. 8). The phylogenetic profiling approach has also been used successfully
to predict the subcellular locations of proteins, based on the observation that
proteins known to localise to a given organelle share characteristic phylogenetic
profiles [62].
Currently, phylogenetic profiling approaches work best for microbes. This is
because a sufficient evolutionary diversity is necessary to be able to identify
distinct phylogenetic profile categories. When comparing very closely related
strains, the profiles tend to be very similar, making it difficult to identify profile
clusters that can be associated with defined functional systems. This will become
even more relevant in the future, when more higher eukaryotic genomes will be
available. Compared to the current phylogenetic diversity of the so far sequenced
genomes of mostly microbial origin, the higher eukaryotic genomes are relatively
similar to each other due to their phylogenetic closeness. Improving the specifity
of the protein families will likely help to improve the resolution of the phylo-
genetic profiles of closely related higher eukaryotes.
Although the phylogenetic profiling approach is considered as a very powerful
method for investigating the functional context of uncharacterised proteins, there
are also limitations in this technique. One difficulty is that the phylogenetic
profiles are often ‘‘noisy’’. This noise is often caused by technical reasons, as the
profiles can only be as good as the predicted genes are. As there are typically some
overlooked genes in a genome, due to shortcomings in the currently available gene
finding programs, the correct profiles can be significantly distorted. Analogously,
gene overprediction can obscure similarities in the phylogenetic profiles. This
difficulty will be gradually resolved, as the gene finding programs constantly
improve in finding the ‘‘correct’’ genes. Another technical difficulty is that the
phylogenetic profiles for some protein classes are not specific enough, as there
are a number of ubiquitous protein motifs causing many proteins to fall into a
few large protein families (e.g., some signalling molecules, such as kinases, or
transporters, see e.g., [64]). Enhancing the algorithms to generate protein families
will likely result in a higher resolution of phylogenetic profiles that is critical
for dissecting distinct transport systems or signalling pathways.

Fusion protein analysis

Analysing fusion patterns of protein domains represents another homology-

independent way to identify functional linkages. In many cases, separated
proteins A and B in organism X are fused in another species Y. When expressed
as a fused protein, the two domains A and B are almost certainly functionally
 Tyrosine Alkaloid
metabolism biosynthesis I

24
L-Tyrosyl-tRNA Puromycin

Spneu
biosynthesis

Mgen
(Tyr)

Hsap
Bsub
Ecoli
L-Tyrosine

L-arogenate 3-(4-Hydroxy
Pentose phosphate Phenylalanine phenyl)pyruvate
Carbon fixation pathway metabolism

L-Phenylalanyl-
tRNA(Phe) L-Phenylalanine Phenylpyruvate
Glycolysis / Prephenate
Gluconeogenesis

D-Erythrose 4 Phosphoenol
pyruvate Indole
phosphate
5-O-(1-Carboxy Histidine 1-(2-Carboxy
Shikimate 3- phenylamino)-1′-
vinyl)-3-phospho deoxy-D-ribulose
phosphate shikimate Chorismate Anthranilate 5′-phosphate
3-Dehydro-3-
deoxy-D-arabino- N-(5-Phospho-D-
heptonate 7- ribosyl) L-Tryptophan
phosphate anthranilate

Ubiquinone
Folate biosynthesis Tryptophan
biosynthesis
metabolism

3-Dehydro
quinate 3-Dehydro
shikimate

Quinate

Fig. 8. Phylogenetic profiling analysis to reconstruct metabolic pathways. Here, the tryptophan biosynthetic pathway with projected
phylogenetic profiles is used to exemplify the similarities in the phylogenetic distribution of the enzymes involved in a common cellular function.
The profiles represent five different genomes, H. sapiens (Hsap), B. subtilis (Bsub), E. coli (Ecoli), M. genitalium (Mgen)and S. pneumoniae
(Spneu), with squares indicating the presence (red) and absence (pink) of the corresponding enzyme. The profiles are overall very similar, and in
particlar along biochemical reaction cascades (see e.g., the reactions framed in yellow). This demonstrates that all enzymes required for a
complete biosynthetic process needs to be encoded in a given genome to guarantee that the ultimate biochemical product can be synthesized.
Thus, phylogenetic profiles can be used to assign uncharacterized gene products to the pathways they are involved. Note that the aminoacyl-
tRNA synthesases (indicated by yellow stars) are encoded by all genomes, including the parasite genome M. genitalium and the mammalian
genome, as the enzymes connecting the aminoacids with the appropriate tRNA are essential for functional protein biosynthesis (PhylosopherTM,
Genedata, Basel).
 25

linked. As the sequences representing A and B are unrelated, this type of

functional relationship is again not detectable via homology-based searches
[40,66,67,79,105].
For instance, GyrA and GyrB, two proteins that are known to interact
physically, are separate subunits in E. coli. These two subunits have been found
to be fused in the yeast topoisomerase II, which shows significant similarities to
GyrA and GyrB [68]. Another example is the human d-pyrroline-5-carboxylate
synthetase, which is made up of E. coli’s g-glutamyl phosphate reductase
and the E. coli glutamate-5-kinase, catalyzing the first two steps of the proline
biosynthesis pathway, but which are not known to interact with each other [79].
Yet another example, representing a fusion of five indpendent protein domains,
is demonstrated in Fig. 9.
Although the fusion protein analysis approach provides valuable insights
into protein–protein interactions and pathways, it can also lead to the pre-
diction of false positive results. One typical error source is that the domain
fusion analysis cannot distinguish between homologs that bind and those that
do not bind. In particular, ‘‘promiscuous’’ domains, such as the signalling
domains SH2 and SH3 cause problems. For instance, for the src molecule the
kinase domain and the SH2 and the SH3 domain are known to interact with
one another (see [79], and references herein). This is certainly not the case
for all proteins harboring an SH2 or SH3 domain, as these motifs are very
commonly used among signalling molecules. Other promiscuous domains, such
as AAA ATPase domains, ATP-binding cassettes, receptor Tyr kinase II
domains, WD repeats and actin binding domains cause similar problems. These
motifs can lead to a considerable false positive rate in a protein fusion analysis. It
could be shown that filtering of the very few promiscuous domains (approx-
imately only 5% of the ProDom domain database) from the predicted
interactions, the number of false positives can be significantly reduced [79].
The dynamics and the strong increase in genome sequences provide promising
perspectives for systematic fusion protein analyses. The more newly sequenced
genomes will become available, the more likely a fusion protein will be found
that will link a domain of unkown function with an already characterised
protein domain. Screening for such functional links indicating direct physical
protein interactions, or at least some functional relationship in a protein com-
plex or a pathway, may lead to the identification of new pathways and protein
complexes.

Neighbourhood correlation analysis

Functional links between proteins can also be revealed by comparing the

local genome organisation between different organisms. If the genes that encode
two different proteins are neighbours on the chromosome across several
genomes, then the proteins tend to be functionally linked. Clearly, this method
is most powerful when applied to prokaryotes, for which spatially clustered
 (a)
Genome X
X1

26
Genome Y
Y1 Y2

known function unknown function

S. cerevisiae
(b) 700000 702500 705000 707500 710000 712500 715000 7175000 720000
YD... PSL10 ARO1 YDR128W
INO2 ECM18 SAC6 FIN1 YDR131C YDR132C ..

ARO1

(c) E. coli
AroB
AroA
AroL
AroK
AroD
AroE
YdiB
3-dehydroquinate EPSP synthase Shikimatekinase Type I Shikimate/
Synthase(PF01761) (PF00275) (PF01202) 3-dehydroquinase quinate
(PF01487) 5-dehydrogenase
(PF01488)

Fig. 9. Fusion protein analysis to predict protein function beyond the homology barrier. (a) Fusion proteins (X1) as key for transferring
functional information from functionally characterized genes (Y1) to non-homologous, uncharacterized genes (Y2); (b) Genome locus of the
yeast ARO1 gene (green arrow) with the domain structure of the ARO1 protein sequence, consisting of five distinct regions (colored bars). (c)
Schematic alignment of the corresponding E. coli proteins to the ARO1 domain structure, with each E. coli protein being similar to a distinct
ARO1 domain. For clarity, the five fused functional motifs encoded by ARO1 are represented by their Pfam motifs (colored horizontal bars).
The yeast ARO1 catalyses five consecutive enzymatic reactions in prechorismate polyaromatic amino acid biosynthesis, while these steps are
catalised in E. coli by individual proteins (PhylosopherTM, Genedata, Basel).
 27

genes that are co-transcribed are common (‘‘operons’’). Still, there are also some
examples that show promise for unraveling functional links within eukaryotes
[40,69,70,105].
Most computational approaches for systematically identifying neighbourhood
relationships start with a definition of gene pairs that are conserved across
multiple genomes. Using these elementary conserved pair relationships, larger
conserved neighbourhood regions can be reconstructed by joining the gene pairs
[71–75]. Figure 10 shows a genome alignment of a number of prokaryotic and
eukaryotic genomes that highlights the conserved nature of gene neighbourhood
relationships. Various variants of neighbourhood analysis algorithms have been
developed to screen large sets of complete genomes for undetected functional
relationships. So far, primarily prokaryotic genomes have been searched for
conserved, similarly organised genome regions representing possible operon
structures.
It has been estimated that it requires at least 10–15 genomes encoding
the functional system of interest to be able to come to a reasonably precise
description of genetic subsystems [71]. Thus, the quality of the functional
predictions is expected to increase, the more genomes are considered within
the analysis.
A generalised version of the algorithm was used to predict entire prokaryotic
transcription units (‘‘operons’’) on a genome-wide scale. Using the well-
investigated E. coli genome as a test case, it could be shown that around 75%
of all known transcription units could be correctly identified [74]. Overall,
about 700 operons were predicted for E. coli, significantly more than known
from current experimental work. This example demonstrates, how computa-
tional strategies such as neighbourhood analyses can contribute to our under-
standing of functional links between proteins, even if no experimental data
is available yet.

Combining evidence

How do the various methods for predicting functional links compare to each
other? Are they complementary, or do they reveal a large overlap and
redundancy? For this, one has to combine the previously discussed methods
for predicting functional links between proteins.
Applying the different methods to the complete set of encoded proteins of
a genome enables the evaluation of the predictive value of these homology-
independent annotation techniques. In a proof-of-concept study, the phylo-
genetic profiling analysis, the domain fusion analysis and the neighbourhood
correlation analysis were applied to Mycoplasma genitalium, a minimal genome
organism that is often used as a standard benchmarking genome in computa-
tional genomics. M. genitalium was compared to 24 other completely sequenced
genomes, most of them prokaryotes. It turned out that significant context
information could be obtained for 50% of all genes, with the neighbourhood
28

(a)

(b)

Fig. 10. Conserved gene organization across phylogenetically diverse genomes. (a) Operon
analysis: Fourteen different eubacterial, archaeal and eukaryotic genomes have been aligned
and oriented with regard to the orthologous genes of the preprotein translocase subunit SecY
(red arrows, for the genome abbreviations see Fig. 7). Orthologous genes besides SecY are shown
using distinct colors for each ortholog class. The conserved arrangement of genes reveals functional
relationships between the different non-homologous genes (see text). In this example, all colored
genes are involved in the cells’ protein biosynthesis machinery (see protein family legend). Note the
differences between Eubacteria and the Archaea A. fulgidus (Aful) and M. janaschii (Mjan).
Although the overall genome organization is similar, whole-gene insertions and deletions of lineage-
specific proteins can be detected; (b) Analysis of syntenic regions: Genome alignment of the human,
mouse and rat FASTK locus (green), encoding the Fas-activated serine/threonine kinase.
Neighbouring genes and their corresponding orthologs are color-coded, revealing a similarly
organized genome structure across human, mouse and rat. Underneath each FASTK gene,
transcript sequences have been aligned, indicating the existence of alternative splicing variants in
humans (PhylosopherTM, Genedata, Basel).

analysis contributing most to this number (40%) [76,77]. In some cases, the
combined approach resulted in clear-cut hypotheses that can be readily tested
in the laboratory. For instance, the genes MG259 and MG347 have been
implicated in translation-related functions in bacteria and mitochondria. Both
gene products are similar to RNA methylases, and MG259 has a signature
 29

similar to nucleic acid adenine methylases. Together with the link to transla-
tion, this indicates that these proteins are involved in methylation of either
tRNA or rRNA [77]. A graphical representation of another combined study [78]
is shown in Fig. 11. This example demonstrates how the yeast prion Sup35,
which acts as a translation release factor in its non-prion state, is functionally
connected to other proteins involved in protein biosynthesis, consistent with
the primary role of Sup35. Sup35 is also linked with protein sorting and a
chaperonin system, which is believed to aid in the folding of newly synthesised
microtubules [78].
This anecdotal evidence clearly underscores the value of computational
methods for predicting functional links. But can the reliability of those pre-
dicted links between non-homologous proteins be further quantified? An attempt
was made to use a ‘‘keyword recovery’’ approach to validate the computer
predictions of functional links. This simple test compares keyword annotations
for both members of a pair of proteins that was predicted to be linked [40,78].
Naturally, this works only for those cases where both proteins are functionally
characterised. An agreement of both keywords indicates that some functional
relationship indeed exists. Applying such systematic consistency checks, signal-
to-noise ratios were determined that range from 3.4 (fusion protein analysis)
to 4.5 (phylogenetic profiling analysis). These numbers have to be compared
with the corresponding values for experimentally determined links, which are
characterised by a signal-to-noise ratio of 8. By combining the different compu-
tational methods the reliability of the predictions can be significantly enhanced,
and signal-to-noise ratios of 7 could be obtained, which is comparable to the
experimentally derived functional relationships [40,78].

Reconstructing pathways II: Regulatory networks

The importance of signalling and regulatory pathways

Protein coding sequences account for only a small fraction of a typical metazoan
genome. In the case of the human genome, less than 2% of the genomic sequence
codes for proteins [37,56]. On the other hand, the unexpectedly low number of
genes that has been identified by sequencing the human and other genomes
cannot account for the physiological and behavioural complexity exhibited by
virtually all organisms from bacteria to higher eukaryotes. It has been argued,
that the complexity of an organism correlates with the number of possible
expression patterns during its life cycle [82]. In fact, the over-representation of
transcription factors compared to other protein classes in higher organisms is
striking and indicates that transcription-factor-rich genomes are likely capable of
having more regulatory states [83]. While yeast encodes only approximately 300
transcription factors, each of the genomes of C. elegans and D. melanogaster
reveal at least 1,000 transcription factors. The human genome most likely
 30
Co-localisation analysis Reconstructed
functional
network

CCT6

Fusion protein analysis

CCT4
Sup45
Sup35
CCT7

CCT5
CCT3

Phylogenetic profiling analysis

CCT2

Fig. 11. Combining various ‘‘beyond homology’’ strategies for a comprehensive investigation of the functional context of uncharacterized
proteins. Here, the yeast prion Sup35, was investigated by applying genome neighbourhood analysis, fusion protein analysis, phylogenetic
profiling analysis and mRNA correlation analysis. A dense network of predicted functional relationships could be reconstructed (yellow circles:
proteins involved in protein synthesis and folding; red: tRNA/mRNA synthesis or splicing, as well as ribosome biogenesis; blue: protein
targeting; white: unknown or other functional category). Thick lines correspond to functional links that have been predicted independently by
more than one computational method, while the dashed lines indicate the few experimental evidences available (modified after [40,78]).
 31

encodes even more than 3,000 transcription factors [37,56,82]. Apparently the
‘‘administrative machinery’’ represented by signalling pathways and transcrip-
tion control had to be significantly expanded in higher organisms compared to
core functions, such as protein biosynthesis, cell-cycle regulation or metabolism
[83]. As there are typically many different transcription factors required to
combine into complexes for influencing transcription activity, a relatively small
number of transcription factors can lead to a enormous number of transcription
factor combinations. These combinations correspond to the large number of
different regulatory states required to respond appropriately to the huge diversity
of environments, developmental stages, tissues and all kinds of combinations of
internal and external stimuli.
At the most basic level, transcription is controlled by binding of transcription
factors to characteristic DNA recognition sequences, so called cis-regulatory
DNA sequences. Transcription factors are typically protein complexes com-
posed of multiple protein subunits. These factors can mediate gene-selective
transcriptional activation or repression of nearby genes. Additionally, there
are large multi-protein RNA polymerase machines required for promoter
recognition and the catalysis of RNA synthesis. Lastly, there are complexes
necessary to remodel and modify chromatin and to assist the transcriptional
apparatus to navigate through chromatin.
Promoters are typically referred to as the regions surrounding the transcrip-
tion start site (TSS) that are able to direct transcription. When several trans-
cription factor binding sites are clustered together, and the mutation of any of
those binding sites affect the up- or down-regulation of a gene in the same
context, the cluster is called a ‘‘regulatory module’’ [86]. If activation of the
module results in induction of the gene, it is termed an ‘‘enhancer’’, otherwise
a ‘‘repressor’’. Enhancers, repressors and promoters are generally known as
‘‘transcription regulatory regions’’.
Currently, it is still difficult to estimate the amount of cis-regulatory DNA
that is essential for binding regulatory proteins. It is known that even simple
organisms like the sea squirt Ciona intestinalis have an estimated 10,000–20,000
tissue-specific enhancers [81]. Another example is the sea urchin Endo16 gene,
probably the best investigated gene in terms of its regulation mechanisms. The
2,700 bp genomic upstream region of Endo16 harbors at least 33 transcription
factor binding sites that are relevant for controlling its transcriptional activity. In
many cases, regulatory regions are far away from the genes that they control.
For instance, tissue specific enhancers can work over distances of 100 kb of
DNA, such as the embryonic enhancers regulating the mouse and human Igf-2
gene which map over 100 kb from the transcription start site [84,85]. A typical
genetic locus of Drosophila contains several enhancers scattered over an average
distance of 10 kb of DNA, while the transcribed DNA typically comprises just 2
to 3 kb. This is in contrast to unicellular eukaryotes such as yeast, where the
regulatory DNA is usually composed of only short sequences of a few hundred
basepairs length, located immediately adjacent to the core promoter.
32

A thorough understanding of gene expression control mechanisms and their

activation through signalling pathways is central for a better understanding
of many diseases and for identifying novel drug targets. Recently, systematic
in silico promoter analyses have also been used for guiding the development of
assays to screen compound libraries for drug candidates [135,151]. Although
there exist a number of experimental techniques to identify physiologically
relevant transcription factor binding sites, the actual experiments are tedious
and the results often difficult to interpret. To improve our understanding of
regulatory principles, it has been suggested to make use of the available complete
genome sequences and the increasing amount of microarray data to system-
atically screen for putative transcription control regions. Indeed, over the last
few years significant progress has been made in reasonably reliably predicting
and characterizing regulatory regions using computer-based approaches.

Models for transcription factor binding sites

Traditional biochemical experiments are still extensively used to characterise

functional transcription factor binding sites. This experimental work has led to
a compilation of a reasonable set of DNA recognition sequences for a number
of transcription factors. For instance, it was shown that a single binding site
of the transcription factor HNF1 (‘‘hepatic nuclear factor 1’’) is represented
by ‘‘AAGTTAATGGATCTG’’ (see e.g., [87,116], and references therein). By
experimentally identifying additional binding sites, a consensus sequence can
often be defined that indicates the most relevant nucleotides for the complex
DNA–protein interaction. Most conveniently, this consensus is represented as
a so called ‘‘position weight matrix’’ (PWM) describing a set of DNA binding
sites for a given factor. The PWM assigns a weight to each possible nucleotide
at each position within the site, reflecting the frequency with which the given
nucleotide occurs at the given position. For instance, the HNF1 binding site
can be represented by a 4  N matrix (with N ¼ 15 corresponding to the width
of the HNF1 binding site model, see Fig. 12). The score of a particular site is
then obtained by summing the corresponding weights [86]. Position weight
matrices have a sound foundation in thermodynamics [87,88,106] and are known
to capture more information as simple consensus sequences, as they represent
DNA-binding energies which are highly predictive of in vitro protein–DNA
interactions.
The main reason for defining PWMs as models for transcription factor
binding sites is of course that these models can be used to in silico screen newly
sequenced DNA for putative binding sites that have not been identified so
far in the laboratory. However, there are two principal difficulties in drawing
conclusions out of PWM-derived TFBS predictions, one related to compiling
databases of transcription factor binding site PWMs, and the other one to the
in silico prediction algorithms.
 33

(a) “AAGTTAATGGATCTG”

14 16 4 0 1 19 20 1 4 13 4 4 13 12 3
(b)
3 0 0 0 0 0 0 0 7 3 1 0 3 1 12
4 3 17 0 0 2 0 0 9 1 3 0 5 2 2
0 2 0 21 20 0 1 20 1 4 13 17 0 6 4

(c)
2
bits

Position
Fig. 12. Mathematical models for transcription factor binding sites. (a) Specific DNA binding
site of the hepatic nuclear factor (HNF1); (b) Position weight matrix (PWM) that has been
developed based on different experimentally identified HNF1 recognition sequences [116];
(c) Sequence logo representation of the PWM highlighting the most relevant nucleotides for bind-
ing the transcription factor HNF1 [143]. Apparently not all residues in a specific binding sequence
(e.g., the sequence shown in (a)) are equally important for the physical DNA–protein interaction.

Little experimental data for defining PWMs

For most transcription factors, there are just not enough experimentally defined
binding sites available to define a reasonable PWM model for the factor’s
binding site. Without having sufficient experimental input, the TFBS PWM
models are normally not optimally defined, and can therefore result in spurious
hits. A second limitation is that many binding sites have only a low-binding
specificity. For instance, it is difficult to come up with a PWM characterizing the
C/EBP factor (CCAAT/enhancer binding protein), as the known binding sites
are very diverse in their sequences [116]. There are many public and commercial
efforts to extract known binding sites from the literature and to compile a
comprehensive set of PWMs. Besides a few very specialised resources, there
are two major sources for TFBS models focusing on higher eukaryotes, the
Transfac data [117] and the Jaspar2 data [118]. However, these data sources
just reveal how limited our current experimental knowledge about individual
transcription factor binding sites is: Transfac harbours not even 200 vertebrate
Other documents randomly have
different content
 Ingluvin, ½ to 2 scruples.[1]
Carbonate Bismuth, ½ to 2 drachms.
Powdered Nux Vomica, 1 to 3 grains.
Mix.

Divide into 12 powders, and put in cachets—one to be given three times a

day after food.
Diet: First simply give milk and Vichy water in equal parts to drink, a small
quantity at a time; also occasionally a little Brand’s beef essence. If sickness is
very persistent, give stomach twelve hours’ absolute rest, during which time
give every four hours a peptonised beef suppository, then try the former diet
again. When solid food is again given, it should at first consist of scraped lean
raw meat, beef, mutton, or veal; for a change, boiled tripe. Thin barley water
is better than plain water to drink.

Chow Chow, Champion Red Craze.

Born June 8th, 1901. Winner of 61 Firsts and Specials and 14 Championships. The property of
Mrs. Scaramanga, 8, Sussex Square, Hyde Park, W.

Thos. Fall, photo.] [face p. 156.

 Inflammation:
Symptoms: Redness, swelling, and tenderness of the part affected; rise of
temperature; thirst; an abscess may form.
Treatment: External Inflammation: Apply following lotion with lint and
bandage, or dabbed on part often:—
Recipe:

Goulard’s Extract of Lead, 1 drachm.

Laudanum, 1 ”
Water to 6 ounces.
Mix.

If the temperature keeps up, say, to 3 or more degs. above normal, it is

pretty certain an abscess is forming, in which case hot linseed-meal poultices
should be applied. To keep the heat in the poultice, cover the outside of it
with a layer of oil silk.
Internal Inflammation: Keep patient quiet; give aperient medicine, as from
two to ten grains[1] of jalapin. If chest affected, put on flannel coat lined with
thermogen wool. When inflammation situated in abdomen, apply heat to part,
as hot flannels, hot-water bottles, or linseed-meal poultices. When
temperature very high, doses of aspirin, from two to ten grains,[1] maybe
given three or four times a day; and if there is much pain, from two to
fifteen[1] drops of laudanum may be given in water the same time.
Diet: Light, as milk, with Benger’s food, stale bread or toast.

Influenza:
Symptoms: Dulness, loss of appetite, rise of temperature, pains in the loins,
quick pulse. There is generally a cough, and bronchitis or pneumonia may
develop. In some cases there is severe diarrhœa, and there may be some
discharge from nose.
Treatment: Place patient in a dry and comfortable room or kennel, the
temperature of which should be kept from 55 to 65 degs. F., according to
whether the dog has been accustomed to live in a house or kennel. If there is
constipation, give a small dose of castor oil; and if the temperature is high,
give from two[1] to ten grains of salicine three or four times a day. When fever
has passed, give salicylate of quinine, from a quarter[1] to a grain, made into
a pill, three times a day.
Diet: Light, whilst there is any fever; but when this has passed,
strengthening food is required, as under-done or raw meat with rice or bread,
also tripe and fish; and if appetite bad, offer some stewed rabbit with rice or
bread.

Inguinal Hernia:
See Hernia.

Insect Bites:
Symptoms: The parts may become much swollen and red, accompanied by
a good deal of irritation.
Treatment: Dab parts with ammoniated quinine; failing this, eau de Cologne
or methylated spirits. When place very tender, the lead and laudanum lotion
as recommended for external inflammation may be used.

Invalid Foods:
When nursing sick dogs the diet is important, as it is so much better to get
the dog to eat something for himself, rather than always pouring food down
its throat. Taking food voluntarily not only does the dog more good, but it is
less worrying to the patient, for when one has to feed with the spoon or
bottle it is necessary to give nourishment very frequently, whereas when it is
taken voluntarily, more, as a rule, is taken at a time, and therefore it is not
necessary to offer food so often. A good meat tea is made with equal parts of
veal, beef, and mutton—say half a pound of each cut up very small, then
slowly stewed for three hours in a pint of water. This should be strained and
given either cold or warm, whichever the dog likes best.
A jelly made from rabbit is also very nourishing, and dogs, as a rule, are
very fond of it, and they will often take it when they refuse everything else. It
should be made as follows:
The whole rabbit should be cut up in small pieces, including the liver; the
leg bones should be cracked, the heads split open, and the whole stewed in a
pint of water for some hours; then strained off, and if there is more than half
a pint reduce it to that quantity, and set aside to cool. This may be given
either cold or hot; a small quantity at a time, as it is, if made as directed, very
strong.
Fish boiled in water, or boiled in milk; and a capital fish soup is made by
stewing white fish, such as whiting, in milk for some time, and then straining
off and giving the soup to drink. Also boiled fish stewed with rice makes a
good food, and the different kinds of fish alone boiled. A food of this kind may
be given to a dog even when he has a fever, especially if he will take it
himself.
Sheep’s brains boiled in milk make an excellent and tempting food.
Calves’ sweet-breads also boiled, or even grilled, dogs are very fond of.
Chickens’ livers grilled make an appetising dish for a dog; and when a dog
is convalescent, and the temperature is normal, he may be even tempted to
eat by offering grilled meat.
Milk of course is one of the best and most nourishing diets, and when the
dog is very weak the white of one egg to every cup of milk is very
strengthening food. For a change, milk, with plasmon added, should be given,
but too much of this latter food must not be given to dogs with a high
temperature.
Sanatogen is a most excellent, strengthening, and easily-digested food.
Dogs will often retain this when they are unable to take any other food.
Benger’s food with milk is also an easily-digested food, as it is partially
predigested. Cases often arise when a dog cannot possibly retain anything in
the stomach, then it is necessary to give nourishment by the rectum, and it is
astonishing what a long while a dog can be kept alive and fairly strong in this
way.
The best kinds of food for giving by rectum are peptonised milk, or
peptonised beef-tea, and peptonised beef suppositories. Burroughs
Wellcome’s are good nutritive suppositories. As to the quantity of milk to be
given per rectum, from one[1] to eight tablespoonfuls, just warmed, every
three or four hours alternatively with one of Burroughs Wellcome’s meat
suppositories.
In giving a nutritive enema, care must be taken to pass it very slowly into
the bowel so as not to excite action, or the enema will be immediately
rejected, and afterwards just raise the hind-quarters a little bit so that the
fluid runs well into the body, and hold the tail down for a few minutes so that
it cannot escape.
The milk can be peptonised with Fairchild’s peptonised powders, which can
be bought at any chemist’s shop.
Brand’s meat essences are excellent foods in cases of stomach disorders.
Benger’s peptonised beef jelly is a very easily digested preparation, and very
useful in cases of severe vomiting.
Raw meat beef-tea, made by soaking for a couple of hours half a pound of
scraped lean raw beef in half a pint of cold water, then stood in front of a fire
to get warm, then straining and squeezing through a coarse tea-cloth. Or the
juice may be pressed from raw meat with one of Dr. Klein’s meat-squeezing
machines. This is very nourishing and easily digested, and dogs are fond of it,
and often will take it voluntarily when refusing other foods.
An excellent combined food for dogs very ill, especially with distemper,
when the patient is very weak, or during convalescence, is made as follows:—
To a breakfast cup of milk, thickened with Benger’s food, add the white of
an egg, a full teaspoonful of invalid bovril, and a dessertspoonful of brandy; of
this give from one[1] to six tablespoonfuls every two or three hours
alternately, with some beef-tea or meat extract.
Messrs. Spratts’ Patent have recently introduced a new food for invalids. It
is a granulated meal, and they call it Invalid Food. It is a most excellent
preparation, and every dog I have offered it to has eaten it with avidity. I
have found it a very useful diet for distemper patients mixed with milk; and I
have given it to puppies just weaned, and they have thriven well on it.
Though this new food is called Invalid Food, it is an excellent preparation
either mixed with milk or soup for small dogs; if meat is required it can easily
be added, as it contains none, but I am told that it contains a special meal,
and that little or no meat is necessary.
The same preparation is put up in the form of biscuits which are crisp
without being hard, and small dogs eat them with pleasure. No doubt all dogs
are better for having something hard to gnaw once a day. It preserves the
teeth, hardens the gums, and assists digestion.
Animal Kreochyle is an excellent extract of meat for use in cases of great
weakness, the result of distemper or from any other disease. It is also an
excellent remedy in stomach disorders accompanied by sickness. Dogs, as a
rule, take Kreochyle very readily, and it is easily digested and assimilated.
 Irritation of Skin:
Symptoms: Constantly scratching, biting, and licking the skin, which when
examined, there is often nothing to be seen. The condition occurs in hot
weather, especially during the shedding of the old coat.
Treatment: Give a sulphur bath made by dissolving one ounce sulphurated
potash in a pail of tepid water; repeat every two or three days. If this does
not give relief, bathe the dog in a warm solution of borax, one tablespoonful
to a gallon of water. Give saline aperient medicine, as Dinneford’s fluid
magnesia, to small dogs, and Epsom salts to large ones. A meat diet is often
beneficial in these cases, but sometimes it increases the irritation; then, of
course, it must be avoided, and other food with green vegetables substituted.

Itch:
See Mange.

Jaundice:
Symptoms: Generally the result of congestion of the liver, caused by chill;
may be due to impaction of the duct with a bile stone, or worms; or the
opening of the duct into the bowel may be stopped by thickening of the bowel
membrane. The attack generally commences with sickness, dullness, loss of
appetite; and the membranes of the mouth and eyes turn yellow, and so does
the skin. There is generally obstinate constipation, and what is passed from
bowel is usually grey or slate colour. The urine is scanty and high-coloured.
Jaundice is also a symptom of organic disease of the liver.
Treatment: When the result of congestion caused by a chill, I have found,
after many years’ experience, that the homœopathic preparations of Nux
Vomica 3 × and Merc. Sol. 3 × act well in these cases. Of the trituration give
from two[1] to ten grains of each every three or four hours. If the bowels do
not operate on the second day, give an enema of from two tablespoonfuls to
half a pint[1] of warm soapy water; repeat daily if necessary. Hot linseed-meal
poultices may be applied to abdomen.
Diet: Mutton broth (in which some green vegetables have been cooked),
with toast or stale bread; milk and Vichy water in equal parts to drink; later
boiled fish or tripe. When the condition is the result of obliteration of the duct,
there is nothing to be done; relief may occur spontaneously, or an operation
may be performed, but it is not recommended. When jaundice is due to
organic disease of the liver, it cannot be cured, but the Nux Vomica and Merc.
Sol. treatment will sometimes give temporary relief. The application of iodine
vasogen over the enlarged liver in these latter cases is often useful in relieving
tension of the organ.

Kennels, How to Disinfect:

After dogs have been suffering from any contagious disease, like distemper
or mange, in a kennel, it is necessary to thoroughly disinfect them, and it is
best done in the following manner:—
In a strong tin dish from half to a pound of sulphur, according to the size of
the kennel, should be placed. A few drops of methylated spirits should be
poured on top of the sulphur, and a light applied, the methylated spirits being
added to make the sulphur ignite more easily. The kennel should then be
made as air-tight as possible by having strips of paper pasted over all the
crevices and around the window-frames, so that the sulphur fumes cannot
escape. The kennel should be kept sealed up for forty-eight hours, after which
the windows and doors may be thrown open so as to thoroughly ventilate the
place, and the following day the whole of the inside kennel should be
thoroughly washed with a strong solution of Pearson’s fluid, 1 in 40. The walls
and ceilings when dry should be afterwards well lime-washed or sulphur-
washed—that is, finely powdered sulphur mixed with water and size, the
same way as whitewash is prepared. A day or two later, when everything is
thoroughly dry, the kennel will be quite fit to receive dogs, or even puppies,
without any fear of infection.

Kidneys (Inflammation of, Acute):

Symptoms: The attack generally comes on suddenly; may arise from severe
chill, but generally the result of stone in the kidneys. There is great pain over
the loins; the dog walks with difficulty and arched back; the temperature rises
3 or 4 degs. above normal; the pulse is quick and full; the urine is high-
coloured and scanty—sometimes it is the colour of blood, and mixed with
mucus and pus; the limbs may swell from dropsy, and the dog is very thirsty
and often frequently sick.
 Treatment: Give saline purgative medicine, as Epsom salts, from twenty
grains to one ounce[1] in warm milk; repeat the next or following day. Also
give every four hours from five to thirty grains[1] of hyposulphite of soda in a
little water. As to food, it should consist principally of milk mixed with equal
parts of Vichy water, and a little boiled fish. If sickness very severe, give from
half to two drops[1] of diluted hydrocyanic acid in a teaspoonful of water
every two, three, or four hours, and ice to lick. After the acute stages have
passed, give tonics, as the ammoniated citrate of iron, from one to five grains,
[1] three times a day, in from one teaspoonful[1] to a tablespoonful of water.

Kidneys (Inflammation of, Chronic):

Symptoms: May be the sequel of an acute attack or a less severe chill, and
sometimes it is caused by gravel. There is tenderness over the loins; the urine
may be tinged with blood, or after passing water, which may look quite clear
and normal, the dog continues to strain, and a few drops of blood escape and
some mucus.
Treatment: A milk diet is important in these cases; it may be given with
rice, tapioca, Force, or bread. Red meat must be avoided, but tripe and fish
may be allowed. Vichy water mixed with equal parts of plain water should be
given to drink; and a course of hyposulphite of soda, as recommended in
acute inflammation, should be given, but only two or three times a day. If
gravel or a stone is suspected, a course of boro-citrate of magnesia with
bicarbonate of potash is advised. The dose of the former is five[1] to thirty
grains, and of the latter from two[1] to fifteen grains mixed with the food
twice a day, and be continued for some time.

Lactation (Defective):
Symptoms: A small supply or total suppression of the mammary secretion.
Treatment: Give a raw meat diet, and stimulate the glands by massage;
also encourage the bitch to drink plenty of fine oatmeal gruel.

Lactation (Excessive):
Symptoms: Excessive secretion of milk. The glands are swollen, hard, and
painful, and the milk often dribbles away. It may occur before the puppies are
born, or just after; and it often happens to a maiden bitch to have a large
secretion of milk, which shows itself about seven or eight weeks after heat,
and which lasts five or six weeks. A maiden bitch in this condition is very
restless and wretched. She appears to be always looking for puppies. She will
walk about with her tail down, crying, and occasionally scratches and rakes at
her bed, and twists round and round as if making a nest.
Treatment: In the first instance, rub the glands with warm salad oil to
soften them, and draw some milk off night and morning. In the second case,
simply draw some of the milk off, and avoid if possible putting anything on the
glands, in case you should injure the puppies or put them off their feed. In
the case of bitches having milk who have not been pregnant, purgative
medicine—castor oil is the best—should be given once or twice a week, and
the glands rubbed with camphorated spirits, or dabbed with a lotion made
with two tablespoonfuls of gin and half a pint of water. When the milk collects
in large quantities as to cause discomfort, it must be drawn off; but this
should be avoided if possible, as it has a tendency to stimulate secretion. Give
the food dry; biscuits are the best.
It is most important to thoroughly attend to maiden bitches when in this
condition, otherwise the milk curdles and becomes hard, and this is the
commencement of mammary tumours.

Larynx (Inflammation of):

Symptoms: A dry, husky cough, the dog after coughing retches as if about
to vomit. If neglected, may go on to bronchitis. It is very contagious.
Treatment: An emetic gives immediate relief, give from one quarter[1] to a
grain of tartar emetic shaken dry on the tongue, also give the following
mixture:—
Recipe:

Liquor Hydrochlorate of Morphia, 2 drachms.

Syrup of Squills, ½ ounce.
Syrup of Tolu, ½ ounce.
Syrup of Lemons, ½ ounce.
Water to 3 ounces.

Doses: From half a teaspoonful to a dessertspoonful[1] two, three, or four

times a day, according to the severity of the cough.
 Also give purgative medicine. Holding the head over hot water, in which a
little Friar’s balsam has been mixed, gives relief, and so do the fumes of
burning cresoline.
The dog should be kept dry and warm; in fact, in one temperature for a few
days.

Lead Poisoning:
Symptoms: Blue line on gums, vomiting, loss of appetite, great thirst,
generally constipation, but there may be diarrhœa. Colicky pains, and in some
cases convulsions, followed by paralysis of the hind legs.
Treatment: Give Epsom salts, from ten grains[1] to one ounce in some
sweetened milk. Also a course of iodide of potassium, from half to four grains
in from one teaspoonful[1] to a tablespoonful of water. For the treatment of
the paralysis, see Paralysis.

Leucorrhœa:
Symptoms: A pale, yellowish discharge from vulva. May occur before heat,
but more often afterwards, and frequently seen after pupping.
Treatment: Syringe with a weak tepid solution of Condy’s fluid, about half a
teaspoonful to half a pint of water. If the discharge is persistent, syringe with
one drachm of powdered burnt alum in half a pint tepid water night and
morning.
Give tonics, as from half[1] to three grains of sulphate of iron (made into a
pill), two or three times a day.

Lice:
See Insect.

Liniment:
The following is a soothing liniment for recent cases of sprains and injuries
to joints, etc.:—
Recipe: The Liniment:
 Tincture Hydrocyamus, 4 drachms.
Methylated Chloroform, 4 drachms.
Spirits of Camphor, 1 ounce.
Soap Liniment, 2 ounces.
Mix.

Apply with friction night and morning.

A stimulating liniment for bronchitis, sore throat, pneumonia, etc.:—
Recipe: The Liniment:

Strong Solution of Ammonia (Liq. Ammon. Fort.), 3 drachms.

Soap Liniment to 4 ounces.
Mix.

Apply with friction night and morning.

A Liniment for Chronic Sprains, Enlarged Joints, and Glands:

Colourless Tincture Iodine, 2 ounces.

Soap Liniment, 2 ounces.
Mix.

Apply with friction once a day.

Lips (Cracked):
Symptoms: The lips along the edges become dry, thick, and crack as the
result of eczema.
Treatment: Paint once or twice a day with sulphurated calcium lotion; when
cracks healed, anoint with boracic ointment. Give cooling medicine, as
bicarbonate of potash and sulphate of magnesia, from two to ten grains of
each,[1] twice a day with food.

Lips (Sore):
Symptoms: Unhealthy-looking ulcers assuming the form of cancer or lupus
occasionally form on the lips.
 Treatment: Clean frequently with a saturated solution of boracic acid, and
paint the ulcer with a twenty-five per cent. solution of chromic acid twice a
week.

Liver (Sluggish):
Symptoms: Indifferent appetite and loss of condition, the coat staring;
bowels constipated, and the motions white or slate colour. Dog vomits in the
morning. Tongue white and coated, and the breath foul. Eyes congested.
Treatment: Give from one to six grains[1] of grey powder, which repeat in a
few days; also give a course of the following:—
Recipe:

Bicarbonate of Soda, 2 drachms.

Tincture Rhubarb, 4 ”
Tincture Nux Vomica, 1 drachm.
Water to 6 ounces.

From one teaspoonful to a tablespoonful[1] three times a day.

Diet: Spratt’s biscuits soaked in soup, with green vegetables added.
 A lever for forcibly opening the jaws in cases of ‘Lock-jaw’

Lock-jaw:
Symptoms: A rare disease, but occurs occasionally, the result of injury to
the head or severe hurt to the eye, and may result from sunstroke. It is
seldom that the whole body as well as the jaw is affected in the dog as in
other animals and people, and, as a rule, it assumes a chronic form. Though
the dog is unable to open his mouth, he is generally able to suck in fluid food,
as milk with eggs and strong beef-tea. The muscles of the head become much
wasted.
Treatment: At first give sedatives, as from two to ten grains[1] each of
hydrated chloral and bromide of potassium in water three or four times a day.
After a time, means must be taken by aid of levers to gently force the jaws
apart. It must be done very gradually by increasing the extent of the opening
a little more each day. The treatment requires to be continued some time to
obtain permanent beneficial results; but directly the dog is able to open his
mouth a little, encourage the gnawing of bones.

Lumbago:
Symptoms: A form of rheumatism affecting the loins. The dog shows signs
of much pain when walking or upon pressure to the parts. Dogs affected with
lumbago often lose all power for a time in the hind legs; in fact, it is the cause
of many cases of paraplegia.
Treatment: See Rheumatism.

Lungs (Inflammation of):

Symptoms: Not a common ailment of dogs, except in cases of distemper or
influenza, but it may result from cold. The attack generally commences with
rigors or shivering; the temperature rises to 103 or 104, in some cases even
higher; pulse is increased in frequency, full and hard. The breathing is quick,
and at each expiration the dog may give a suppressed grunt. The chest is
tender on pressure. If the ear is put to the chest, crepitation will be detected;
and as the disease advances, the parts of the lungs affected become quite
dull, and there are no sounds to be heard except the air passing in and out of
the large tubes. At the commencement of the attack, the dog may have a
husky cough, but it generally stops after a day or so, to commence again
later. The dog is generally off his food, and the eyes are congested.
Treatment: Place the dog in a jacket lined with thermogen wool, or apply
hot linseed-meal poultices right round chest. If no better on second day, apply
a blister to front of chest; the liquor epipasticus is as good as anything. The
hair must first be cut off closely over the part, and the blister rubbed into the
skin for five minutes. If the skin is not blistered the next day, rub a little red
blister ointment into the place. Care must be taken that the dog is not allowed
to lick the blisters, as they are poisonous. For medicine give from one[1] to
ten grains of phenacetin every six hours to reduce the temperature, but it
must not be continued for long; brandy may also be given in small quantities
often. If the dog becomes very weak and the pulse feeble, give every four or
six hours from two[1] to ten drops of Tincture Digitalis, with from one[1] to
five drops of Tincture Nux Vomica, in water. In bad cases the inhalation of
oxygen relieves the distressing breathing. It is important to keep the bowels
open; in fact, it is generally a good plan to give a purge at first. The diet must
be light, and consist of milk, beef-tea, meat extracts, and such like food.

Lupus:
See Lips (Sore).

Maggots:
 Dogs with a long coat kept in a dirty state sometimes become infested with
maggots, especially in those parts near the tail.
Treatment: Wash daily for a week with Pearsons fluid diluted eighty times
with warm water. Afterwards dry and comb out all mats. In very bad cases it
is best to cut the hair off short.

Mange (Sarcoptic or Common):

Symptoms: This disease, which is very contagious from one dog to another,
and readily caught by people, is due to a small insect. The complaint once
contracted soon spreads more or less all over the body, but the most favourite
spots for it to attack is the skin around the eyes, the outside of the ears, the
elbows, and the outside of the hind legs, as well as the skin covering the
abdomen, and underneath parts of the chest. Small red spots like flea-bites
appear where the insect burrows into the epidermis, and the acrid matter
which they excrete sets up intolerant irritation, causing the dog to constantly
scratch, breaking the coat, which is now very brittle, and leaving bare
patches, besides injuring the skin and creating sores which dry and scab. If
there is any doubt about the case, the skin should be scraped where bad with
a knife, and the scrapings examined under a microscope, and if the disease is
mange the parasite will be found.
Treatment: The disease is easily cured, and the specific remedy is
powdered sulphur, one part mixed with eight parts of vegetable oil, which
should be thoroughly well rubbed all over the dog every four days for three
times; three or four days after the last dressing, the dog may be washed.
It is important to thoroughly disinfect the kennels by fumigation, and well
washing the walls and floors with a strong solution of Pearson’s fluid; also all
the collars, leads, combs, and brushes used for the dog, should be disinfected
by baking or soaking in a solution of Pearson’s fluid.

Mange (Follicular):
A skin disease confined principally to puppies, though adult dogs do
occasionally contract it. It is not contagious to people.
Symptoms: It is a slowly progressive disease, and may commence with a
single circular bare patch, about the size of a shilling, on the face or side of
nose. The disease is, of course, not confined to the head, as the first sign may
appear on some part of the body or one of the legs. The patch is generally of
a dirty grey colour, and upon which will be found a number of reddish pimples
or elevations of the skin, somewhat larger than those seen in ordinary mange;
some contain a blood-coloured fluid, others ordinary pus, or matter tinged
with blood, which is easily evacuated by squeezing. This fluid contains the
parasite, which looks, when examined under the microscope, like a small silk-
worm.
As time goes on, the original patch increases in size and others form, the
pimples break, one running into another, and unhealthy-looking sores result.
When these wounds heal, the skin has a dry, corrugated appearance, and
little excrescences of skin are formed, and the hair does not always grow
again.
The skin in follicular mange generally turns a dark greyish-blue or black
colour, and the disease is called by some people “black mange”.
Treatment: It is a most unsatisfactory disease to treat, for often after
months of hard work the dog is no better, but on the contrary is much worse,
the disease having progressed in spite of everything. If the patient is not a
valuable dog, and the attack a bad one, it is much better to destroy him at
once; however, when it is decided to give the dog a chance, treatment should
commence by having the dog, if a long-coated one, clipped all over, so that
the sores may be got at, and then dress him all over with the following:—
Recipe:

Black Sulphur, 2 ounces.

Kerosene, 4 ounces.
Olive Oil, 4 ounces.
Cocoa-nut Oil, 4 ounces.
Castor Oil, 4 ounces.
Wright’s Solution of Coal Tar, 1 ounce.
Well mix.

With this, dress the dog all over once a week, but before each dressing
have him thoroughly washed, using Cook’s 3% mercurial soap, and carefully
dry before applying the dressing. To the spots apply Naphthol Beta ointment
daily. If this does not heal them, then dab on the sores twice a day peroxide
of hydrogen (20 volumes). For a change, the sores may be dressed with the
following:—
Recipe:
 Oil of Cade, }
Methylated Spirits, } Equal parts mixed together.
Green Soft Soap, }

In treating follicular mange a change of dressing to the sores is necessary.

During the treatment an occasional course of arsenic often does good; give
from one[1] to five drops of liquor arsenicalis in water after food; gradually
increase the dose to from two[1] to ten drops. Continue the medicine for
about three weeks at a time, and after a week’s interval give it again as
before.

Marasmus:
Symptoms: Loss of condition, colour, flesh, and strength.
Treatment: A raw meat diet with tonics, or the following pills:—
Recipe:

Reduced Iron, 12 to 48 grains.[1]

Arsenic, ⅟₁₂th to ¼th ”
Extract Gentian, 6 to 20 ”
Mix.

Divide into 12 pills. One twice a day after food.

Cod-liver oil is also advised after food, from half a teaspoonful[1] to a
tablespoonful twice a day.

Mastitis (Inflammation of Breast):

Symptoms: This complaint is not uncommon in bitches when nursing
puppies. One or more of the milk-glands may be affected. The breast is
swollen, red, and very tender. The bitch is off her feed, and there is often a
good deal of fever. An abscess generally forms, which should be lanced as
soon as it becomes soft and points.
Treatment: Hot water fomentations are the best, as it is difficult to apply
poultices when a bitch is nursing puppies. If the temperature is very high,
from two[1] to ten grains of salicylate of soda may be given three or four
times a day.
 After the abscess has been opened, or has broken, tonics as from a quarter
of a grain[1] to one grain of sulphate of quinine should be given three times a
day.

Meningitis (Inflammation of the Membranes of the

Brain):
Symptoms: Occasionally occurs in young puppies, particularly when
suffering from worms, but is more often the result of a protracted attack of
distemper.
In young puppies, the patient rolls its head from side to side, is constantly
whining and crying, has convulsions, and the head is burning hot. When the
result of distemper, convulsions are always present, the dog champs the jaws,
emitting a quantity of frothy saliva. These are what are called distemper fits.
The eyes are congested, and there are generally two, three, or more degrees
of fever. The patient is frequently crying or whining, denoting evident signs of
pain. As a rule, in these cases the patient becomes greatly emaciated and
very weak—in fact, paralysed; and often, as the inflammation extends to the
spinal cord, which it frequently does, chorea or St. Vitis’s dance sets in, which
is practically incurable.
Treatment: When occurring in young puppies, worm medicine should
always be given; and to allay the pain and stop the convulsions, small doses
of bromide of potassium be given—from two[1] to five grains in a little milk
three or four times a day. Ice may also be applied to the head in a sheep’s
bladder. When the convulsions have quite stopped, small doses of cod-liver oil
do good. When meningitis follows distemper, if the attack is a bad one, there
is little hope of recovery, and as a result the proper course to pursue is to
have the dog mercifully put out of his misery. However, people seldom will
agree to this without, as they say, giving the dog a chance, and when the
case is not a bad one recovery is not impossible, though the chances are
greatly against it, for usually it is a progressive disease. If the bowels are not
operating (but there is generally diarrhœa) give a dose of castor oil, and
commence a course of bromide of strontia at once. Give from two[1] to ten
grains, or even fifteen grains, if the dog is a very big one, every three, four, or
six hours, according to the severity of the case. Apply an ice-bag to the head
for a quarter of an hour or longer at a time, and put a seton in the back of the
neck, just behind the ears, which dress with turpentine ointment. Keep the
dog absolutely quiet in a dark room, and feed entirely on a liquid diet—as
milk, with white of egg; Benger’s food or toast; and some beef-tea. Brand’s
essence or rabbit soup may be given occasionally, but those latter foods are
more stimulating than nourishing. The milk foods mentioned are perfect food,
and will support the dog an indefinite time. For the treatment of chorea and
paralysis, see the articles on those subjects.
The bromide must be continued for some time, even though the
convulsions stop, but given in less quantities.

Milk:
It is not generally known how much stronger or more nourishing a bitch’s
milk is than cow’s milk, and when I had some samples analysed I was
surprised myself. Below I give the analysis of a fair sample of each kind:—

Cow’s. Bitch’s.
Water, 87·4 66·3
Butter, 4·0 14·8
Sugar and Soluble Salts, 5·0 2·9
Caseine and Insoluble Salts, 3·6 16·0

Practically a bitch’s milk is nearly three times as strong as a cow’s, and yet
people, when giving the latter to puppies, often dilute it with one-third of
water, and thus add to its weakness. Consequently a much larger quantity has
necessarily to be given to a puppy for it to derive sufficient nourishment to
sustain it, and as a result the stomach is over-distended. Indigestion follows,
and the puppies do not thrive. All breeders must have noticed how often
puppies go back directly artificial food is given them, even before being
entirely weaned, and this, no doubt, is in a great measure due to improper
food in overloading the stomach with a quantity of waste and superfluous
fluid. Even when cow’s milk is given pure, nearly three times the quantity in
bulk ought to be given than if bitch’s milk is used, and it is thus seen at a
glance how a puppy’s stomach, which is naturally small, is over-distended. A
dog’s digestive organs are arranged to receive food in a concentrated form—
as, for instance, dogs in nature eat only meat, and until they get this food
they have nothing else but the dam’s milk, which is also a concentrated food.
When about five or six weeks old the mother goes out in search of food, and
comes back with her stomach loaded, which, after a time, when it is partially
digested, she ejects by vomiting, and this the puppies eat, and thus they get
naturally partially-digested food suitable for a puppy’s delicate stomach.
 To treat cow’s milk so as to make it as like bitch’s milk as possible, to every
three-quarters pint of the former add two and a half ounces of cream—that is,
about three tablespoonfuls—two and a quarter ounces of plasmon, and five
ounces of water. First mix the plasmon with the water, add the milk and slowly
boil in an enamel saucepan for two minutes, then add the cream when cold
and well mix.
When feeding puppies with artificial bitches’ milk, it must be remembered
that only a small quantity is necessary—for instance, about the third part one
would give of cow’s milk.

Milk (Defective Secretion):

See Lactation (Defective).

Milk (Excessive Secretion):

See Lactation (Excessive).

Mouth:
See Gumboil, Lips, Teeth, Toothache.

How to bandage a dog’s head

 Mumps:
Symptoms: Swelling of side of face, and just below the ear, which is very
tender. Dogs suffering from this complaint are generally rather feverish. There
is a loss of appetite and an excessive flow of saliva.
Treatment: Give a dose of purgative medicine and cover swelling over with
a piece of spongiopyline, or a pad of cotton-wool covered over with oil-silk,
and apply a bandage or put on a cap. Later, when the swelling has
commenced to go down, give the following tonic pills:—
Recipe:

Reduced Iron, 6 to 30 grains.[1]

Sulphate of Quinine, 4 to 12 ”
Common Mass. sufficient.
Mix.

Divide into 12 pills. One to be given twice a day one hour after food.

Muscles (Wasting):
Symptoms: There may be general wasting of all the muscles of the body,
the result of some long illness, as distemper; or there may be wasting of the
muscles of one or more limbs, the result of want of use, as in cases of injury;
or some disease of the spinal cord, causing paralysis, and occasionally wasting
of the muscles of one or both sides of the head and face, the result of some
injury to the head.
Treatment: When the result of illness, no special treatment is required
beyond giving tonics, as cod-liver oil, for when the patient gets about again
the muscles will fill up as before. When one or more limbs are affected as the
result of disuse from lameness or paralysis, the limbs should be massaged
and galvanised. When the head is affected, the case is often very obstinate,
and sometimes the muscles never regain their normal size, but the best
treatment is massage galvanism, and encourage the dog to gnaw big bones.

Muzzling a Dog:
 A dog muzzled with tape

How to muzzle a bulldog with tape

The piece over the forehead is necessary in a bulldog to draw the

tape off the short nose to permit easy breathing

The best way to muzzle a dog effectually, so that he cannot bite, is with a
piece of strong tape which should be passed over the top of the nose, tied
twice under the chin, and then the ends should be carried behind the ears,
and again tied tightly.
Care must be taken in muzzling a bulldog, or you may interfere with his
breathing. Therefore, after tying behind the head, one end of the tape should
be passed under the tape which crosses the top of the nose, and this part
pulled well up and tied with the other end of the tape over the forehead, so
as to remove any pressure from the top of the nose. (See Illustration.)

Nails (Cutting):
This cannot be safely done with scissors, but proper nail-clippers should be
used. When the nails are white, it is an easy matter, as the quick can be seen,
and the nail may be cut off within the sixteenth of an inch of the quick; but
when the nail is black, the quick cannot be seen, then the first one must be
cautiously cut a little at a time. The dog will soon wince when the nippers are
getting near the sensitive part. When one is done, it is an easy matter doing
the rest, as they may be cut off in the same proportion as the first.

Nails (Dew-Claws):
The dew-claws or side-nails should always be removed from the back legs a
few days after birth, by being cut off close to the limb with a pair of sharp
scissors. I always consider it would be a good plan if it was a custom to
remove the dew-claws from the forelegs of all dogs also; of course, in some
cases, as in fox-terriers and others, it is done.

Nails (Injuries, etc.):

Symptoms: Dogs with defective action wear their nails away (more
particularly those on the forelegs) to the quick, which causes lameness. Some
dogs’ nails are very brittle, and they crack and split to the quick, causing
lameness. Toe-nails of dogs who do not have sufficient exercise turn over as
they grow, and the point becomes embedded into the pad, causing swelling,
suppuration, and great pain. The dew-claws of dogs who do not get the
chance of digging often grow to a considerable extent, which weakens and
makes them liable to break, which often happens, and the quick becomes
exposed and bleeds, causing the dog to go lame as if he had a broken leg. If
they do not break, they grow, entering the pad as previously described.
 Treatment: In cases of a dog wearing the claws away through defective
action, as a rule, treatment is of little use; but if the dog is made to exercise
on grass land, the nails will grow to an ordinary extent, and the dog walk
much better, but the relief is only maintained whilst he is not allowed to run
on hard roads. When the defective action of the limbs causing undue wearing
of the nails is due to partial general paralysis—as the sequel of distemper, for
instance—then it is only temporary, and will pass off as the patient improves
in health and strength. To remedy the cracking and splitting of the nails is
often a difficult matter. In some cases benefit is derived by the application of
tar ointment, which should be well rubbed into the nails twice a day. The
frequent application of glycerine is also a good remedy. In very bad cases,
especially when only one nail is affected, it may be extracted.
When a nail has grown too long and injured the pad, the nail should be cut
close to the quick without making it bleed; and when there is any festering,
hot bread or linseed-meal poultices should be applied for a day or so.
Afterwards the wound should be dressed with boracic ointment, and the foot
kept in a canvas bag for a few days.
When a dew-claw becomes broken, it is nearly always through the quick. In
such cases it is necessary to remove the claw by extraction with a pair of
tooth forceps, and not by cutting, or the quick will be injured, and the nail will
grow deformed. Afterwards anoint the raw surface with boracic ointment,
cover over with a pad of boracic wool and bandage.
 Four Celebrated Stud Dogs: Buck Stone, British Stone, Dick Stone, and Rex Stone.
The property of Walter Jeffries, 28, Grove Park, Denmark Hill, S.E.

[face p. 188.

Nasal Catarrh:
Symptoms: Sneezing; a watery discharge from nose, followed by a semi-
purulent discharge and coughing; but unlike distemper, the attack is seldom
accompanied by fever and with little loss of condition, and as a rule, the dog
is better in a few days.
Treatment: Put in a dry, warm room of a temperature about 60 degs. F.;
give a dose of aperient medicine as from two[1] to ten grains of jalapin.
The next day, when this has worked off, commence the following pills:—
Recipe:

Salicylate of Quinine, 2 to 12 grains.[1]

Solution Arsenic, 3 to 12 minims.
Solution Sulphate Atropine, ½ to 2 drops.
Extract Gentian, 5 to 20 grains.
Mix.

Divide into 12 pills—one to be taken every four hours.

When the discharge has ceased, and the appetite has become normal, if
the cough continues troublesome, give from half[1] a teaspoonful to a
tablespoonful of cod-liver oil two or three times a day.

Nasal Parasites:
Symptoms: Very occasionally dogs suffer from a worm in one of the nostrils
called Pentastioma Tænoides. It is a worm varying in size from one and a half
to two inches long, and about half an inch wide in the centre, tapering slightly
at each end. It develops at the back of the nose, and whilst growing appears
to cause no discomfort to its host; but when it commences to move, it induces
a good deal of sneezing and a discharge of mucus from the nasal passage
from the side where it is lodged. The dog eventually dislodges it during a
violent attack of sneezing.
 I have never seen more than one in a case, and the dog seems comfortable
directly it has got rid of its guest.

Nasal Polypus:
Symptoms: A growth with a narrow neck which forms in the nasal passage
causing irritation, sneezing and snorting, a purulent discharge which may
sometimes be tinged with blood.
Treatment: Consists in removing the growth by ligature if it can be got at,
or by an ecraseur or a snare; but it is a surgical case, and a veterinary
surgeon should be consulted.

Necrosis (Diseased Bone):

Symptoms: Generally the result of some injury. The parts are swollen,
inflamed, and very painful, and generally an abscess forms, which, if not
lanced, bursts, discharging a thick, creamy and often offensive-smelling pus.
The bone, which is generally dead, lies exposed at the bottom of the wound,
which does not heal permanently until the dead bone comes away.
Treatment: At first, hot linseed-meal poultices dusted over with powdered
charcoal are to be applied, and continued for some days after the abscess is
open. Then lint saturated with carbolic oil, one in forty, is to be placed over
the wound, and kept in place with a bandage. As a rule, after some time the
dead bone exfoliates and comes away of itself, and the wound then soon
heals, but in many cases, an operation is necessary to remove the dead bone.

Nephritis (Inflammation of the Kidneys):

Symptoms: Generally commences suddenly with an attack of shivering and
a rise of temperature, there being often three or four degrees of fever
accompanied by a rapid pulse. There is pain in the back, vomiting, scanty and
high-coloured urine, or blood may be mixed with the water. In some acute
cases it is quite suppressed, and the dog then suffers from uræmic poisoning.
It may arise from a severe chill, but oftener from a stone in the kidneys.
Treatment: Give milk and Vichy water in equal parts to drink, also water to
drink, to which has been added a teaspoonful of cream of tartar to every pint.
Administer a dose of Epsom salts, from one scruple[1] to one ounce, dissolved
in warm water, but given cold. If vomiting persistent, give from half[1] to two
drops of diluted hydrocyanic acid in a teaspoonful of water, and ice to lick.
After the acute stage has passed, give from one[1] to five grains of
ammoniated citrate of iron, three times a day in water.

Nettlerash:
Symptoms: Generally arises suddenly, and often the result of a chill, as, for
instance, a dog plunging into cold water when hot after exercise. When once
a dog has had an attack, he is liable to a recurrence, and then indigestion will
often induce it. The skin becomes suddenly nodulated or swollen in patches,
the hair standing on end. The ears may be affected and become half an inch
thick; the head is often attacked, and the dog is temporarily blind because the
eyelids are swollen. Sometimes one or more legs are the seat of the trouble,
and become three or four times their normal size; in other cases, large
patches appear about the body. As a rule, it passes away almost as quickly as
it comes.
Treatment: Keep dog quiet and warm whilst the attack is on, and give from
half to two drops[1] of liquor arsenicalis (P. B.) in a little water every two or
three hours. Also give a good dose of aperient medicine. For a few days after
a dog has had nettlerash, he should be kept on a light milk diet, and given a
course of arsenic, as from one to eight drops[1] of liquor arsenicalis in water,
three times a day after food.

Neuralgia:
Symptoms: Dogs are affected with a form of neuralgia that attacks the
muscles and nerves of the neck and shoulders. The pain, which is very acute,
comes on quite suddenly, and the dog cries with it. The head is drawn into
the body, as it were. The muscles are swollen, and very tense. There is a
difficulty in walking; and when it is attempted, it is done very slowly and
stiffly. The attack may last a couple of hours, sometimes longer, and then
slowly passes off, and the patient may appear quite well; when all at once it
may come on again as bad as before, and so keeps on for days, and
sometimes weeks, unless some suitable treatment is adopted.
Treatment: The dog must be kept very quiet in these cases, as very often
any sudden movement induces an attack. Give from two to ten grains[1] of
salicylate of soda three times a day, made into a pill; and if the attacks of pain
are very acute, also give two or three times a day, injected under the skin,
from one-twentieth to the eighth of a grain[1] of hydrochlorate of morphia,
with from ⅟₃₀₀th to ⅟₁₀₀th of a grain[1] of atropine, in from five[1] to ten
minims of distilled water. At first these injections induce vomiting, but it soon
ceases, and the treatment is a specific for the disease.

Nipples (Sore):
Symptoms: The nipples become swollen, inflamed, and cracked very often
after a bitch has been nursing puppies some weeks, so much so that at last
she will not allow the puppies to suckle.
Treatment: Wash the parts two or three times a day with boracic lotion, and
anoint with boracic ointment after drying.

Nose (Bleeding):
Symptoms: May be due to injury of the head or face. It often occurs in
cases of pneumonia, and it is a symptom of a polypus in the nose or
ulceration of the membrane lining the nasal passages.
Treatment: When slight, no particular treatment is required; but if
persistent, the cause should be ascertained, and if possible removed. When
severe, as the result of some injury, ice may be held to the bridge of the
nose; and if this does not stop it, then pour or inject about fifteen drops of
the solution of adrenalin diluted four times with water into one or both
nostrils. It is not advisable to plug the nostrils, as it distresses a dog to
breathe through his mouth.

Nursing:
The first thing to be considered in nursing a sick dog is proper quarters for
the patient to live in, for in all cases of serious illnesses he should not be
allowed to run loose about a house and out of doors when he likes. If a house
dog, he should be put in a good, well-ventilated room, with the temperature
kept at as near 60 degs. F. as possible. Of course, in very hot weather that
cannot be done, but as much air as possible must be given by keeping the
windows wide open during the summer months. In winter or cold weather,
the temperature of the room should be kept up to 60 degs. F. by means of
artificial heat—an ordinary fire is best; failing this, a paraffin stove—avoid a
gas stove if possible. Of course, with dogs who are in the habit of living out of
doors it is different; but even with them, dry, large, well-lighted and
ventilated, comfortable quarters, free from draughts, are absolutely necessary
if the patient is to have a fair chance, and the temperature of the place should
be kept up to 55 degs. F. A loose box in a stable that is kept clean makes a
capital place; but unless the stable is kept very clean, it smells of ammonia,
which is fatal to a dog suffering from distemper, because pneumonia and bad
eyes are sure to develop.
Sick dogs should always be kept separate. It is a great mistake to put three
or four together.
As to diet for patients, see article on Invalid Food; when they will not take
food voluntarily, a small quantity, varying according to size of dog, must be
given often, about every two hours or oftener, day and night. It is useless to
feed a dog well for sixteen hours, and to leave him to chance for eight. It is
often during the night, when the system is at its lowest, that a little good
nourishing food, with some stimulant, is most wanted; and it is this attention
that saves the patient in many cases.
The preparation of food is most important. It should either be done by
one’s self, or under the personal superintendence of a responsible person. All
milk food should be made fresh three times a day, and any that is left over
should be thrown away. All feeding utensils must be kept scrupulously clean,
and the spoon, bottle, or feeding-cup that is used for food or medicine should
be washed and dried immediately after being used, ready for next time. The
cooking utensils also must be kept scrupulously clean. If these things are not
attended to, diarrhœa and sickness result, and the patient is weakened, and
perhaps has a relapse. Also keep the sick dog scrupulously clean. Each time
after feeding clean his lips with a little weak solution of Condy’s fluid on a pad
of cotton-wool—which should be thrown away immediately after using—and
dry with soft cloth; also cleanse the fundament and prepuce once a day with
the same preparation, and keep the eyes free of discharge with a little weak
boracic acid lotion, and also clean the teeth with a weak solution of
permanganate of potash. This is very important.
As well as attending to the patient, the room or kennel requires frequent
attention. Do not make the air stuffy with strong disinfectants, but it is a good
plan to sprinkle the floor, whether a kennel or room, with pine sawdust, and if
the flooring be wood, to cover it over with sheets of old newspapers, which
may be covered with sawdust, and then all evacuation can be carried away
and burnt, for when a dog is very ill he ought not to be allowed to go out.
There are some dogs who are so clean that they will not make themselves
comfortable in a room, and it becomes absolutely necessary to let them out
rather than make them worse by keeping them in. But a dog may be kept for
some hours, twelve or even sixteen, to see if he will not give way. Once he
has relieved himself in the room, and finds he is not scolded, he gains
confidence, and is not so particular in the future. To make an obstinate dog
do what is necessary in a room, especially when the weather is bad, and
when perhaps it would be fatal to let him out, I give either a dose of aperient
medicine or an enema, which invariably has the desired effect.
Do not always be fidgeting an invalid. Do what is necessary, and then leave
him alone.
Take the temperature regularly three times a day, at the same time each
day, and keep a record on a chart; also, if you can, at the same time count
the pulse and the number of respirations per minute, and record them for
reference. In all cases of severe illness, it is a good plan to put the dog in a
flannel jacket, as depicted in the illustration.

A coat covering the hind parts of a dog—useful after operations

upon the abdomen

The points to be remembered are:—

1. Dry, well-ventilated, light quarters of a proper temperature.
2. Cleanliness of patient.
3. Cleanliness of the surroundings.
Welcome to our website – the ideal destination for book lovers and
knowledge seekers. With a mission to inspire endlessly, we offer a
vast collection of books, ranging from classic literary works to
specialized publications, self-development books, and children's
literature. Each book is a new journey of discovery, expanding
knowledge and enriching the soul of the reade

Our website is not just a platform for buying books, but a bridge
connecting readers to the timeless values of culture and wisdom. With
an elegant, user-friendly interface and an intelligent search system,
we are committed to providing a quick and convenient shopping
experience. Additionally, our special promotions and home delivery
services ensure that you save time and fully enjoy the joy of reading.

Let us accompany you on the journey of exploring knowledge and

personal growth!

ebookultra.com