Sociedade Brasileira de Genetica Medica

Access Provided by:

The Online Metabolic and Molecular Bases of Inherited Disease

The Population Genetics of PAH

Judith R. Kidd; Kenneth K. Kidd

DOI: 10.1036/ommbid.100

PREFACE
When clinical geneticists think about phenylalanine hydroxylase (PAH), they think about mutations in the gene causing phenylketonuria (PKU) and the
other metabolic diseases of phenylalanine hydroxylase deficiency, their diagnosis, and their treatment. When biochemical geneticists think about PAH,
they think about metabolic pathways, enzyme activity levels, and metabolic substrates and products. When population geneticists think about PAH,
they think about allele frequencies of disease­causing mutations and normal polymorphisms in different populations, the patterns of these variations
in the populations, and what these frequencies and patterns mean in terms of disease gene and, more broadly, human evolution. We are population
geneticists, and thus we will describe our studies of normal variation in the PAH gene in populations around the world.

INTRODUCTION
Soon after the PAH cDNA was cloned, it was used as a probe to detect restriction fragment length polymorphisms (RFLPs) by Southern blotting (Lidsky
et al,45; Woo et al,63). The linkage between PKU and PAH was, of course, very strong. In the effort to discover the mutation(s) leading to the
hyperphenylalanemias, searching for a nonrandom association of disease with alleles of normal polymorphisms [linkage disequilibrium (LD)] in the
PAH gene was a strategy to consider. The use of this strategy has now reached epidemic proportions in human genetics, especially for complex
diseases (Capon et al,6; Horikawa et al,32; Kim et al,41; Myers et al,47; Sawcer et al,52; Scapoli et al,53; Vaessen et al,60), but often has low power unless
there is strong prior support for a specific gene. In the case of PAH and PKU there was no question that the disorder is biochemically a deficiency of the
enzyme and genetically inherited at (or very close to) the PAH gene. This is the very circumstance in which LD should be at its most powerful.
Nevertheless, the simple version of this strategy did not work in the case of PKU because there was not a single etiologically crucial mutation, but,
rather, many different PKU­causing mutations distributed across the approximately 90 kb of the gene. Recourse was to haplotypes. A haplotype is the
combination of alleles at two or more polymorphisms close enough to one another on a chromosome that the alleles at the polymorphisms are usually
transmitted through generations as an intact stretch of DNA. That is, recombination does not disrupt or scramble the combinations of alleles in the
haplotype regions. The term is derived from haploid genotype (Ceppellini et al,7) and can refer to a specific combination of the variants at multiple sites
or can refer to the entire system.

Although PKU showed no significant association with any single RFLP, in a sample of Danish families certain haplotypes were found to be
disproportionately associated with the disease­causing mutations. This finding led to the identification of the first two specific PKU mutations (DeLella
et al,15; DeLella et al,14) and confirmed the value of PAH haplotypes as a tool for understanding the population and mutational genetics of PKU
(Kidd,39). Thenceforth haplotypes at PAH were used to give an initial indication of which mutation(s) were present in a patient or family with PKU or
phenylalaninemia states, and also to signal that a new (i.e., previously undescribed) mutation was present (Apold et al,2; Baric et al,4; Daiger et al,1989a;
Daiger et al,1989b; Dianzani et al,19; Hertzberg et al,30; Jaruzelska et al,34; Konecki, Lichter­Konecki,43; Kozak et al,44; Stuhrmann et al,55; Svensson et
al,56; Zygulska et al66). Polymorphisms, mutations, and haplotypes of the PAH region had become well characterized in patients with PKU by 1996
(Nowacki et al,48; Woo,62). Summaries of PAH mutations have been published (e.g., Eisensmith et al,22; Konecki, Lichter­Konecki,43), and an up­to­date
compendium of PAH mutations and background haplotypes is maintained at the PAHdb website (http://www.pahdb.mcgill.ca) (Nowacki et al, 1998).

PKU is particularly common in northern Europeans (Bickel et al,5) and in Yemenite Jews (Avigad et al,3), approaching an incidence of 1 in 5000 in both
populations. PKU is much less common in most other populations; therefore, interest in PAH in other parts of the world has tended to lag behind that
in Europe and the United States. Rarely were the non­PKU haplotypes studied in populations of non­European origin except in those families with PKU.
In a study of a small number of non­PKU African Americans (Hofman et al,31), researchers found more haplotypes than in Europe or Asia. In Japanese
Downloaded 2022­2­28 8:15 P Your IP is 189.6.248.104
and Chinese (Daiger et al,1989b) and in a sample from several Polynesian groups (Hertzberg et al,30), on the other hand, the number of normal
Page(non­
The Population Genetics of PAH, Judith R. Kidd; Kenneth K. Kidd 1 / 21
PKU) haplotypes was reduced relative to that in Europe. These studies also concluded that,
©2022 McGraw Hill. All Rights Reserved. Terms of Use • Privacy Policy • Notice • Accessibilityas in Europeans, LD exists among the common
polymorphisms across both ends of the gene at distances of 22 and 31 kb, and is much weaker (and generally not statistically significant) between the
markers at opposite ends of the gene (Chakraborty et al,8; Daiger et al,1989a). This pattern of the molecular extent of linkage disequilibrium agrees
compendium of PAH mutations and background haplotypes is maintained at the PAHdb website (http://www.pahdb.mcgill.ca) (Nowacki et al, 1998).
Sociedade Brasileira de Genetica Medica
PKU is particularly common in northern Europeans (Bickel et al,5) and in Yemenite Jews (Avigad et al,3), approaching anby:
Access Provided incidence of 1 in 5000 in both
populations. PKU is much less common in most other populations; therefore, interest in PAH in other parts of the world has tended to lag behind that
in Europe and the United States. Rarely were the non­PKU haplotypes studied in populations of non­European origin except in those families with PKU.
In a study of a small number of non­PKU African Americans (Hofman et al,31), researchers found more haplotypes than in Europe or Asia. In Japanese
and Chinese (Daiger et al,1989b) and in a sample from several Polynesian groups (Hertzberg et al,30), on the other hand, the number of normal (non­
PKU) haplotypes was reduced relative to that in Europe. These studies also concluded that, as in Europeans, LD exists among the common
polymorphisms across both ends of the gene at distances of 22 and 31 kb, and is much weaker (and generally not statistically significant) between the
markers at opposite ends of the gene (Chakraborty et al,8; Daiger et al,1989a). This pattern of the molecular extent of linkage disequilibrium agrees
with the early findings of Jorde and colleagues36 that LD generally does not extend more than 50­60 kb in populations of European origin as well as
with the more recent studies that also suggest that LD usually does not extend beyond ∼50 kb (e.g., Gabriel et al,26; Reich et al,50).

Our own studies of LD in multiple populations have found that LD among normal polymorphisms can differ dramatically among populations from
different regions of the world (Chattopadhyay et al, in press; DeMille et al,16; Kidd et al,33; Kidd et al,38; Tishkoff et al,57; Tishkoff et al,58;). These studies
have shown that, in general, Africans have much lower values of LD than do Europeans or European Americans, a finding supported by several other
studies (Frisse et al,25; Gabriel et al,26; Jorde et al,37; Kittles et al,42; Lonjou et al,46; Purandare et al,49; Reich et al,50). Specifically at PAH, we have shown
that overall LD is statistically significant at p <.01 in every population, but pairwise LD extends farther in Native American populations than elsewhere in
the world (Kidd et al,38). Although in that study there was a hint of less LD in Africans, it was still significant across the same parts of the locus that
showed significant LD in Europeans. Here we shall summarize our findings in a more global sample of populations and additional polymorphisms
across the locus. The primary objective is to understand better the evolution of the normal PAH gene and to define the background haplotypes in
diverse populations upon which clinically relevant mutations arose. We have collected data on seven polymorphisms including six of the "classic"
single­nucleotide polymorphisms (SNPs) defined initially as Southern blot RFLPs (Dworniczak et al,1991a; Dworniczak et al,1991b; Goltsov et al,28;
Lidsky et al,1985b; Wedemeyer et al,20), part of the somewhat larger set of eight RFLPs used to define haplotypes based on Southern blotting with the
cDNA (Woo,62).

Using the sequence from GenBank and unpublished protocols from Dr. Randy Eisensmith, we have confirmed or developed polymerase chain reaction
(PCR) protocols for six of the classic SNPs and an exon 7 AluI restriction­site SNP. The map of the gene with locations of these polymorphisms is shown
in Fig. 101­1. Here we present the haplotype frequencies for PAH based on the seven SNPs.

Fig. 101­1

Schematic map of the PAH locus indicating exons and sites of the single­nucleotide polymorphisms studied. For comparison, locations of two other
commonly studied polymorphisms, the tetranucleotide short tandem repeat polymorphism (STRP) in intron 3 and the EcoRV restriction fragment
length polymorphism downstream of the gene, are also indicated. Drawn approximately to scale, as indicated, with 66.98 kb between the BglII and
XmnI sites and 34.02 kb between the PvuIIb and EcoRI sites.

METHODS
Populations

The population samples we have studied are listed in Table 101­1 along with their unique identifiers (UIDs) in ALFRED (http://alfred.med.yale.edu). A
description of each sample can be found in ALFRED indexed by its UID. All samples were collected with both approval from the appropriate institutional
review boards and informed consent from the participants. The DNA in this study was purified, by means of standard phenol­chloroform extraction
and ethanol precipitation (Sambrook et al,51), from Epstein­Barr virus­transformed lymphoblastoid cell lines (Anderson, Gusella,1). These samples
were collected for purposes unrelated to PKU, and no information is available on whether any relative has PKU. We assume that all chromosomes
contain a normal PAH allele.
Downloaded 2022­2­28 8:15 P Your IP is 189.6.248.104
The
SNPPopulation
Typing Genetics of PAH, Judith R. Kidd; Kenneth K. Kidd Page 2 / 21
©2022 McGraw Hill. All Rights Reserved. Terms of Use • Privacy Policy • Notice • Accessibility
We have used the genomic sequence in GenBank to determine the molecular locations of and distances among the various polymorphisms (cf.
accessions AC069227 and AC079952). The seven SNP sites and their UIDs in ALFRED are given in Table 101­2.
The population samples we have studied are listed in Table 101­1 along with their unique identifiers (UIDs) in ALFRED (http://alfred.med.yale.edu). A
description of each sample can be found in ALFRED indexed by its UID. All samples were collected with bothSociedade Brasileira
approval from de Genetica
the appropriate Medica
institutional
review boards and informed consent from the participants. The DNA in this study was purified, by means ofAccess
standard phenol­chloroform
Provided by: extraction
and ethanol precipitation (Sambrook et al,51), from Epstein­Barr virus­transformed lymphoblastoid cell lines (Anderson, Gusella,1). These samples
were collected for purposes unrelated to PKU, and no information is available on whether any relative has PKU. We assume that all chromosomes
contain a normal PAH allele.

SNP Typing

We have used the genomic sequence in GenBank to determine the molecular locations of and distances among the various polymorphisms (cf.
accessions AC069227 and AC079952). The seven SNP sites and their UIDs in ALFRED are given in Table 101­2.

Ancestral Alleles

Segments around each of the SNPs were amplified from DNA of several other apes, including at least two other species: chimpanzee, bonobo, gorilla,
and orangutan. These PCR products were sequenced by standard cycle sequencing using fluor­labeled dideoxynucleotides and analyses on an Applied
Biosystems 377.

Data Management and Simple Statistics

The databases and programs as described by Kidd and colleagues38 were used. The database for managing data is PhenoDB (Cheung et al,10). While
population structure or nonrandom sampling would be other possible explanations for significant deviation from Hardy­Weinberg (H­W) ratios, they
are not considered likely for the population samples studied here since these same samples have been studied for multiple other loci and such
deviations have not been found generally and nearly all loci would be affected by structure in a population.

Haplotype Frequency Estimation

The population samples we study consist largely of unrelated individuals; therefore, family data cannot be used to determine phase in multiply
heterozygous individuals. We used maximum­likelihood estimates of haplotype frequencies and calculated the standard errors (jackknife method)
from the multisite marker­typing data, implemented in either HAPLO (Hawley, Kidd,29), or the more elaborated HAPLO/P (Zhao et al,65), both programs
that implement the EM algorithm (Dempster et al,17). The EM algorithm iteratively allocates individuals proportionately to all possible genotypes based
on the current haplotype frequencies and then re­estimates haplotype frequencies by "counting" the numbers of occurrences of each haplotype.
Obviously, the "occurrences" are not whole numbers when there are multiply heterozygous individuals probabilistically allocated to different
genotypes. However, wherever an individual is homozygous or heterozygous at only one site, only one genotype is possible and that allocation does
not vary from iteration to iteration; such individuals constrain the process. The iterations converge to a maximum likelihood estimate of haplotype
frequencies. Several studies have indicated that haplotype frequencies estimated by the EM algorithm are generally accurate (Fallin, Shork,23; Tishkoff
et al,59; Zhang et al,64). Note that the genotypes of multiply heterozygous individuals are not determined except probabilistically. Thus, if four
haplotypes (++, +−, −+, −−) are observed among unambiguous individuals in the sample, each individual that is doubly heterozygous, +−,+−, will be
either ++/−− or +−/−+ with probabilities dependent on the relative frequencies in the sample of the two genotypes.

Linkage Disequilibrium Analyses

LD in its most general sense is a deviation of haplotype frequencies from those expected if alleles at the various sites occur randomly on
chromosomes. Various statistics can be used to quantify that deviation (Devlin, Risch,18) and various approaches can be used to determine whether the
observed deviation from randomness is more than would be expected by chance, i.e., the statistical significance of the deviation. A classic statistic for
quantifying LD is the statistic D and the standardized value D′ that varies from −1 to +1 with 0 being equivalent to no LD, i.e., random association of
alleles at two sites. There are several problems with D′. One problem is that D′ is defined for haplotypes of two biallelic sites whereas there are now
regions of DNA with multiple polymorphic sites including multiallelic sites. The other problem is that statistical significance is not correlated with D′ so
that different haplotype systems, all with D′ = 1 (the maximum possible LD), can be a mixture of highly significant observations and nonsignificant
observations. A better statistic is r2 which is more closely correlated with significance. However, it is also limited to pairwise comparisons of sites. For
large numbers of sites within small genomic regions, as is the case for PAH, such pairwise comparisons do not capture all the nonrandomness. Zhao
and colleagues65 proposed a measure of overall nonrandomness, ξ, for multiple sites with multiple alleles. A permutation­based method to estimate ξ
was proposed and is implemented in HAPLO/P. The permutation process also provides a test of statistical significance. In addition to the multiple­site
application, ξ can also be used for pairwise analyses; in those analyses it is highly correlated with r2 (Zhao et al, unpublished results).

One variant of the ξ statistic is the segment test. It measures the nonrandomness across a segment of DNA using sets of markers on either side. In this
Downloaded 2022­2­28 8:15 P Your IP is 189.6.248.104
test, discussed in more detail by Zhao and colleagues65 and Kidd and colleagues,38 the typing results for individuals are held constant for the markers
The Population Genetics of PAH, Judith R. Kidd; Kenneth K. Kidd Page 3 / 21
within a set and only permuted between the two sets. Thus, the measure controls for any
©2022 McGraw Hill. All Rights Reserved. Terms of Use • Privacy Policy • Notice • AccessibilityLD within each set. Use of more than one marker on each side
of a segment compensates for variation in heterozygosity and integrates information from markers that are not immediately adjacent but still show LD.
large numbers of sites within small genomic regions, as is the case for PAH, such pairwise comparisons do not capture all the nonrandomness. Zhao
Sociedade Brasileira de Genetica Medica
and colleagues65 proposed a measure of overall nonrandomness, ξ, for multiple sites with multiple alleles. A permutation­based method to estimate ξ
Access Provided by:
was proposed and is implemented in HAPLO/P. The permutation process also provides a test of statistical significance. In addition to the multiple­site
application, ξ can also be used for pairwise analyses; in those analyses it is highly correlated with r2 (Zhao et al, unpublished results).

One variant of the ξ statistic is the segment test. It measures the nonrandomness across a segment of DNA using sets of markers on either side. In this
test, discussed in more detail by Zhao and colleagues65 and Kidd and colleagues,38 the typing results for individuals are held constant for the markers
within a set and only permuted between the two sets. Thus, the measure controls for any LD within each set. Use of more than one marker on each side
of a segment compensates for variation in heterozygosity and integrates information from markers that are not immediately adjacent but still show LD.

RESULTS
Allele and Haplotype Frequencies

The allele frequencies for each of the sites are given in ALFRED under the site UIDs given in Table 101­2. (In addition to the samples we have studied,
ALFRED will in the future contain data from some additional population samples studied by others.) The frequencies of common haplotypes for the 7­
SNP haplotype system are given in Table 101­3. Haplotypes that are never seen at a frequency of ≥10% in any population are pooled into the residual
class. The full tables for all 7­SNP haplotype frequencies, including jackknife standard errors when available, are present in ALFRED [under the UIDs in
Table 101­2 (as of July 2003, the data were being entered)]. To make the results more accessible, the frequencies of the 15 most common 7­SNP
haplotypes are shown in Fig. 101­2 as stacked­bar histograms of regional averages. This representation emphasizes the global variation in haplotype
frequencies but ignores still significant variation within each region, especially among the African populations as can be seen in Table 101­3 and
Figures 101­3, 4, 5, 6, and 7. Fig. 101­2 makes two points: multiple haplotypes exist in Africa with none at a very high frequency (elaborated in Fig. 101­
3), and East Asian populations have one particular haplotype at an unusually high frequency (elaborated in Fig. 101­5).

Fig. 101­2

The average haplotype frequencies by geographic region represented as stacked­bar histograms. The region and number of distinct population
samples in the average are given on the left of each bar. The length of each colored segment represents the average frequency in that region of the
corresponding haplotype as indicated at the right of the figure. The alleles are indicated for the sites in order from 5′ to 3′ (“1” for the site­absent allele,
“2” for the site­present allele). Specifically graphed are the 15 most common haplotypes, those that occur in at least 1 population at a frequency of at
least 10%. Frequencies of all haplotypes that never attain this frequency in any of the sampled populations are summed together in the residual class
represented by the gray bar. Figures 101­3 through 7 show the individual population frequencies underlying these averages.

Fig. 101­3

The haplotype frequencies for each population in the African average in Fig. 101­2. The same graphical conventions are used.

Downloaded 2022­2­28 8:15 P Your IP is 189.6.248.104

The Population Genetics of PAH, Judith R. Kidd; Kenneth K. Kidd Page 4 / 21
©2022 McGraw Hill. All Rights Reserved. Terms of Use • Privacy Policy • Notice • Accessibility
 Sociedade Brasileira de Genetica Medica
Fig. 101­3
Access Provided by:

The haplotype frequencies for each population in the African average in Fig. 101­2. The same graphical conventions are used.

Fig. 101­4

The haplotype frequencies for each population in the Southwest Asian and European averages in Fig. 101­2. The same graphical conventions are used.

Fig. 101­5

The haplotype frequencies for each population in the East Asian average in Fig. 101­2. The same graphical conventions are used.

Downloaded 2022­2­28 8:15 P Your IP is 189.6.248.104

The Population Genetics of PAH, Judith R. Kidd; Kenneth K. Kidd Page 5 / 21
©2022 McGraw Hill. All Rights Reserved. Terms of Use • Privacy Policy • Notice • Accessibility

Fig. 101­6
 Sociedade Brasileira de Genetica Medica
Fig. 101­5
Access Provided by:

The haplotype frequencies for each population in the East Asian average in Fig. 101­2. The same graphical conventions are used.

Fig. 101­6

The haplotype frequencies for each population in the Northwest Asian (Komi Zyrian and Khanty), Siberian (Yakut), and Pacific (Micronesian and Nasioi
Melanesian) averages in Fig. 101­2. The same graphical conventions are used.

Fig. 101­7

The haplotype frequencies for each population in the North and South American averages in Fig. 101­2. The same graphical conventions are used.

Downloaded 2022­2­28 8:15 P Your IP is 189.6.248.104

The Population Genetics of PAH, Judith R. Kidd; Kenneth K. Kidd Page 6 / 21
©2022 McGraw Hill. All Rights Reserved. Terms of Use • Privacy Policy • Notice • Accessibility
 Sociedade Brasileira de Genetica Medica
Fig. 101­7
Access Provided by:

The haplotype frequencies for each population in the North and South American averages in Fig. 101­2. The same graphical conventions are used.

Overall LD and Distribution of LD across the Gene

The seven­site coefficient of overall nonrandomness is graphed in Fig. 101­8. Figures 101­9a and 9b are contour plots of the pairwise ξ values for all
sites from the 5′ end (e.g., BglII­PvuIIa, BglII­PvuIIb, BglII­EcoRI) and the statistical significance of those values, respectively. Figures 101­10a and 10b are
contour plots of the pairwise ξ values of all sites from the 3′ end of the gene (e.g., XmnI­MspI, XmnI­AluI, XmnI­EcoRI) and the statistical significance of
those values, respectively.

Fig. 101­8

The overall nonrandomness of the 7 PAH single­nucleotide polymorphisms measured as ξ for each of the 38 populations studied. All values are
significantly different from 0 (p <.001). Populations are arranged by geographic region: Africa, Southwest Asia, Europe, Northwest Asia, East Asia,
Pacific, Northeast Siberia, North America, and South America. All populations are indigenous to their regions except for African Americans and
European Americans.

Fig. 101­9

Contour plots of linkage disequilibrium across the PAH locus for each of the 38 populations studied. A. The nonrandomness, measured as pairwise ξ,
for increasing distance from the BglII site, from bottom to top. Each vertical column represents a separate population ordered from left to right as in
Fig. 101­8. B. The
Downloaded statistical
2022­2­28 significance
8:15 P Your IPofisthe ξ values, plotted as in A.
189.6.248.104
The Population Genetics of PAH, Judith R. Kidd; Kenneth K. Kidd Page 7 / 21
©2022 McGraw Hill. All Rights Reserved. Terms of Use • Privacy Policy • Notice • Accessibility
Fig. 101­9
Sociedade Brasileira de Genetica Medica
Access Provided by:

Contour plots of linkage disequilibrium across the PAH locus for each of the 38 populations studied. A. The nonrandomness, measured as pairwise ξ,
for increasing distance from the BglII site, from bottom to top. Each vertical column represents a separate population ordered from left to right as in
Fig. 101­8. B. The statistical significance of the ξ values, plotted as in A.

Fig. 101­10

Contour plots of linkage disequilibrium across the PAH locus for each of the 38 populations studied. A. The nonrandomness, measured as pairwise ξ,
for increasing distance from the XmnI site, from bottom to top. Each vertical column represents a separate population ordered from left to right as in
Fig. 101­8. B. The statistical significance of the ξ values, plotted as in A.

Downloaded 2022­2­28 8:15 P Your IP is 189.6.248.104

The Population Genetics of PAH, Judith R. Kidd; Kenneth K. Kidd Page 8 / 21
©2022 McGraw Hill. All Rights Reserved. Terms of Use • Privacy Policy • Notice • Accessibility
Fig. 101­10

Sociedade Brasileira de Genetica Medica

Contour plots of linkage disequilibrium across the PAH locus for each of the 38 populations studied. A. The nonrandomness, measured as pairwise ξ,
Access Provided by:
for increasing distance from the XmnI site, from bottom to top. Each vertical column represents a separate population ordered from left to right as in
Fig. 101­8. B. The statistical significance of the ξ values, plotted as in A.

Segment Tests

As a means of determining how LD is distributed across the locus, we have calculated segment tests across the three large segments between the
PvuIIa site and the MspI site. This corresponds to the large internal segment in the study by Kidd and colleagues,38 across which there is essentially no
significant LD except in Native American populations. In this study we have tested for LD by the segment test across the three large spans in this locus.
The three spans are from PvuIIa to PvuIIb, from PvuIIb to EcoRI, and between EcoRI and MspI. (For these analyses we have omitted the AluI site <1 kb
from the MspI site.) The results for the three tests are shown in Fig. 101­11.

Fig. 101­11

The segment test ξ values for nonrandomness for the three larger intervals across the middle of the PAH gene for each of the 38 populations studied.
For these analyses the AluI site was omitted. The segments analyzed are indicated by the sets of sites on either side using initial letters for the sites:
(BPPE)(MX), green diamonds; (BP)(PEMX), light blue triangles; (BPP)(EMX), dark blue squares. Essentially all values >0.3 are significantly different from 0
(p <.01; usually p <.001 except for those only slightly greater than 0.3); values <0.2 are not different from 0. Populations are arranged as in Fig. 101­8.

Downloaded 2022­2­28 8:15 P Your IP is 189.6.248.104

The Population Genetics of PAH, Judith R. Kidd; Kenneth K. Kidd Page 9 / 21
©2022 McGraw Hill. All Rights Reserved. Terms of Use • Privacy Policy • Notice • Accessibility
The segment test ξ values for nonrandomness for the three larger intervals across the middle of the PAH gene for each of the 38 populations studied.
Sociedade Brasileira de Genetica Medica
For these analyses the AluI site was omitted. The segments analyzed are indicated by the sets of sites on either side using initial letters for the sites:
Access Provided by:
(BPPE)(MX), green diamonds; (BP)(PEMX), light blue triangles; (BPP)(EMX), dark blue squares. Essentially all values >0.3 are significantly different from 0
(p <.01; usually p <.001 except for those only slightly greater than 0.3); values <0.2 are not different from 0. Populations are arranged as in Fig. 101­8.

Ancestral Alleles

The ancestral states for four of the sites have been previously reported (Iyenger et al,33; Kidd et al,38): site present (2) for BglII, site absent (1) for PvuIIa,
site present (2) for MspI, and site absent (1) for XmnI. Here we report that the ancestral alleles are site absent (1) for PvuIIb, site present (2) for EcoRI,
and site absent (1) for AluI. Whereas site­present alleles must have exactly the same sequence, site­absent alleles could differ from the site consensus
in different ways. However, we found that for each of the site­absent ancestral alleles, the sequence in the other great apes corresponded to the
human site­absent allele. Thus, the ancestral haplotype is 2112121, a haplotype found in most African populations and in several non­African
populations, though usually at much lower frequencies.

DISCUSSION
Allele and Haplotype Frequencies

Considerable variation in allele frequencies exists among the populations for each of the individual sites. This variation can be seen graphically in
ALFRED. Of greater interest, however, is the variation in haplotype frequencies shown in Figures 101­2, 3, 4, 5, 6, and 7. Simple inspection of the
regional averages in Fig. 101­2 shows the tremendous haplotype frequency variation among populations from different parts of the world. The
haplotypes are ordered left to right corresponding to top to bottom in the legend according to global average frequency, not their evolutionary
relationship. The bright red haplotype (2112121) is the ancestral haplotype. It is obvious that by region it is most frequent in Africa and very uncommon
in some other parts of the world. One of the most obvious aspects of this figure is that African populations have more common haplotypes at roughly
comparable frequencies as well as more very rare haplotypes (the residual class) than do populations in other regions of the world. Note also that the
globally most frequent haplotype (1211121, light orange) is quite uncommon in East Asia. The most common haplotype in the Americas (2122121, dark
green) is virtually absent in East Asia and the two Pacific populations. Figures 101­3, 4, 5, 6, and 7 make these points even more obvious when the
frequencies in the individual populations are represented graphically.

Figures 101­3, 4, 5, 6, and 7, showing the individual population frequencies, make another point very strongly: the regional averages are reasonably
representative for Southwest Asia, Europe, East Asia, and the Americas but not at all representative for the individual African populations. In most
other regions of the world, geographically similar populations are similar in haplotype frequencies, but in Africa the frequencies of individual
haplotypes have considerable variation. What all African populations share is that they all have multiple common haplotypes. Comparison of the
African­American frequencies in Fig. 101­3 with the African average in Fig. 101­2 shows an interesting similarity. It seems likely that this similarity
reflects the fact that the African­American population is not just an admixture of a single African population with European populations but is primarily
a mixture of different African ancestries.

The other graphic representations of haplotype frequencies in individual populations emphasize the general regional similarities among populations
in contrast to what is seen in Africa, as noted above, but also show that on occasion a haplotype will exist in a moderate frequency in one population in
Downloaded 2022­2­28 8:15 P Your IP is 189.6.248.104
a region and be undetected or very rare in other populations from the region. For example, the haplotype 2112111 (dark orange) exists in Page a frequency
The Population Genetics of PAH, Judith R. Kidd; Kenneth K. Kidd 10 / 21
of nearly 20% in the sample of Samaritans, but it is not seen or is relatively uncommon in
©2022 McGraw Hill. All Rights Reserved. Terms of Use • Privacy Policy • Notice • Accessibilityalmost all other populations including the African
populations. Similarly, the haplotype 1212112 (light green) exists at a frequency of >10% in the Chuvash and the Ashkenazi Jews; it is seen much less
frequently in other European and African populations but exists at frequencies of nearly 15% in some of the Asian populations. Finally, we note an
African­American frequencies in Fig. 101­3 with the African average in Fig. 101­2 shows an interesting similarity. It seems likely that this similarity
Sociedade Brasileira de Genetica Medica
reflects the fact that the African­American population is not just an admixture of a single African population with European populations but is primarily
Access Provided by:
a mixture of different African ancestries.

The other graphic representations of haplotype frequencies in individual populations emphasize the general regional similarities among populations
in contrast to what is seen in Africa, as noted above, but also show that on occasion a haplotype will exist in a moderate frequency in one population in
a region and be undetected or very rare in other populations from the region. For example, the haplotype 2112111 (dark orange) exists in a frequency
of nearly 20% in the sample of Samaritans, but it is not seen or is relatively uncommon in almost all other populations including the African
populations. Similarly, the haplotype 1212112 (light green) exists at a frequency of >10% in the Chuvash and the Ashkenazi Jews; it is seen much less
frequently in other European and African populations but exists at frequencies of nearly 15% in some of the Asian populations. Finally, we note an
unusual haplotype, 2122212 (plum), that we have seen at appreciable frequencies in only three populations: the San Francisco Chinese, the Khanty
from Northwest Asia, and the Ticuna from the Amazon basin. While the haplotype may have a common origin in the Asian populations, its
"reappearance" in South America may be attributable to a more recent origin by recombination and large amounts of drift in the quite inbred Ticuna.

The haplotype frequencies represented in this series of figures are the basic biological data. It is often useful, however, to consider some derived
statistical representations, most specifically linkage disequilibrium.

Linkage Disequilibrium

Figures 101­9 and 10 show how LD decays with distance across the PAH gene for each population. In Figures 101­9a and 10a, the populations are
ordered across the bottom of the contour graphs as they are in Fig. 101­8, as well as vertically from the top to bottom in Figures 101­2, 3, 4, 5, 6, and 7.
The distance on the vertical axis is the distance from the BglII site and from the XmnI site, respectively, to each of the other six sites studied along the
gene. The colors in the contour plots represent the ξ LD values, increasing from 0.0 to 0.3 and above as the colors range from blue to red. The axes in
Figures 101­9b and 10b are the same as in Figures 101­9a and 10a, but now the color represents significance of the ξ values shown in Figures 101­9a
and 10a. All of the ξ values for the distance/population combinations shown in shades of blue are not statistically significant.

With 7 sites, there are 21 pairwise segments, and with 38 populations, we calculated 798 LD values. Figures 101­12a and 12b summarize these values by
population and segment length (a) and their statistical significance (b). In this summary of LD, we are neglecting the linear relationship of the segments
on the chromosome and some interesting points emerge. Of the 21 pairwise segments, 7 are shorter than 10 kb in length (0.9, 1.8, 2.6, 5.4, 6.4, 7.4, and
9.2 kb), and across these segments almost all populations are in strong LD usually p =.001. There are five segments longer than 60 kb (60.2, 62.0, 64.2,
65.7, and 67.5 kb), and across these segments almost no population is in significant LD (p ≤.05) except the Micronesians (p ≤.05), nearly all Native
American groups (p ≤.001), and others only sporadically. Across the 9 intermediate­length segments (19, 24.5, 34.0, 40.5, 41.4, 43.2, 51.4 56.6, and 58.8)
LD is distributed sporadically. In these segments, Europeans, European Americans, and native Americans often have significant LD Europeans and
European Americans often at the p ≤.05 or p ≤.01 level of significance and Native Americans at the p ≤.001 level of significance. East Asians, Africans,
and African Americans are rarely in LD at any accepted level of significance. In the case of East Asians, heterozygosity is often low, making LD
impossible to measure. The populations that do not align with any geographic cluster in this set are the Yakut, Micronesians, Nasioi, Komi Zyrian, and
Khanty. The Yakut have no significant LD (p ≤.05) in the segments longer than 41.2 kb; the Micronesians have no significant LD in the segments between
20 and 52 kb, but are in significant LD above (p ≤.05) and below (p ≤.01) this length; the Nasioi exhibit strong and significant (p ≤.001) LD up to a
segment length of 43 kb, and no statistically significant LD in longer segments; and the Komi Zyrian and Khanty show vary rare statistical significance (p
≤.05) at distances longer than 35 kb.

Fig. 101­12

Contour plots of linkage disequilibrium across the PAH locus for each of the 38 populations studied. A. The nonrandomness, measured as pairwise ξ,
for all pairs of markers, by distance from bottom to top. Each vertical column represents a separate population ordered from left to right as in Fig. 101­
8. B. The statistical significance of the ξ values, plotted as in A.

Downloaded 2022­2­28 8:15 P Your IP is 189.6.248.104

The Population Genetics of PAH, Judith R. Kidd; Kenneth K. Kidd Page 11 / 21
©2022 McGraw Hill. All Rights Reserved. Terms of Use • Privacy Policy • Notice • Accessibility
Fig. 101­12

Sociedade Brasileira de Genetica Medica

Contour plots of linkage disequilibrium across the PAH locus for each of the 38 populations studied. A. The nonrandomness, measured as pairwise ξ,
Access Provided by:
for all pairs of markers, by distance from bottom to top. Each vertical column represents a separate population ordered from left to right as in Fig. 101­
8. B. The statistical significance of the ξ values, plotted as in A.

Overall LD

The overall LD values are significant (p ≤.001) for all populations. This pattern builds on that found for four of the seven sites (Kidd et al,38) in that, on
average, African populations had the lowest values, East Asian populations the next highest values, and European and Native American populations
the highest values. Since all populations have at least one short segment with highly significant LD, it is not surprising that overall LD is also highly
significant. Based on examination of the pairwise LD values (just discussed), and the distances between the sites, it seemed more reasonable to
examine linkage disequilibrium separately for the three 5′ sites and the four 3′ sites as well as the LD across the segment between them.

For every adjacent pair in the 5′ end of the gene, most combinations (haplotypes) occur. For the two PvuII sites, one of the possible combinations (2­2)
is infrequent and scattered among populations, but it does occur at a frequency as high as 8­9%. Interestingly, the 1­1 haplotype combination at the
first two sites does not occur in the Komi Zyrian, Khanty, Atayal, or Cheyenne; the 2­2 combination does not occur in the Samaritans, Adygei, Russians,
Khanty, Hakka, Japanese, Ami, Atayal, or Mexican Pima; the 2­1 combination does not occur in Ethiopian Jews. Only the 1­2 combination occurs in all
populations tested. LD (D′) is stronger between PvuII(a) and PvuII(b) (18 kb apart) than between BglII and PvuII(a) (5.4 kb apart). All three of the sites are
required to define the four common haplotypes.

For every adjacent pair in the 3′ end of the gene, all four possible allele combinations occurred, though usually one of the combinations is quite
infrequent. Even for the closest two markers (AluI and MspI at 0.9 kb), the four­gamete test clearly shows that ancestral recombination must have
Downloaded 2022­2­28 8:15 P Your IP is 189.6.248.104
occurred. At thisGenetics
The Population end of theofgene
PAH,there areR.
Judith four major
Kidd; haplotypes
Kenneth requiring three SNPs (EcoRI, AluI, and MspI) to distinguish among them. Page 12 / 21
K. Kidd
©2022 McGraw Hill. All Rights Reserved. Terms of Use • Privacy Policy • Notice • Accessibility
Other Polymorphisms across PAH
Khanty, Hakka, Japanese, Ami, Atayal, or Mexican Pima; the 2­1 combination does not occur in Ethiopian Jews. Only the 1­2 combination occurs in all
Sociedade Brasileira de Genetica Medica
populations tested. LD (D′) is stronger between PvuII(a) and PvuII(b) (18 kb apart) than between BglII and PvuII(a) (5.4 kb apart). All three of the sites are
Access Provided by:
required to define the four common haplotypes.

For every adjacent pair in the 3′ end of the gene, all four possible allele combinations occurred, though usually one of the combinations is quite
infrequent. Even for the closest two markers (AluI and MspI at 0.9 kb), the four­gamete test clearly shows that ancestral recombination must have
occurred. At this end of the gene there are four major haplotypes requiring three SNPs (EcoRI, AluI, and MspI) to distinguish among them.

Other Polymorphisms across PAH

The EcoRV site is also an SNP. However, it is located within an Alu element ∼10.5 kb downstream of the end of the gene. We were unable to design a
reliable PCR­based typing protocol. We have also not yet studied the variable number of tandem repeats (VNTR) just downstream of the gene (the
"HindIII RFLP") or analyzed data for the tetranucleotide short tandem repeat polymorphism (STRP) in intron 3 just 4 kb downstream from the PvuIIb
site. However, we have begun to study some of the numerous other SNPs more recently identified and typed by other means such as TaqMan. As data
on those individual sites are collected, they will become available through ALFRED.

Segment Tests of LD

Three segment tests across the middle interval are reported in this study (Fig. 101­11). These intervals are between PvuII(a) and PvuII(b), between
PvuII(b) and EcoRI, and between EcoRI and MspI. In the first of these segment tests, [BglII and PvuII(a)] are permuted against [PvuII(b), EcoRI, MstI, and
XmnI]; in the second segment test, [BglII, PvuII(a), and PvuII(b)] are permuted against (EcoRI, MspI, and XmnI); finally, in the third segment test, [BglII,
PvuII(a), PvuII(b), and EcoRI] are permuted against (MspI and XmnI). The first segment test spans 19 kb; only the Biaka, most East Asian populations,
and the Nasioi are not in strong and statistically significant LD (p ≤.01). Across the second segment, spanning almost 34 kb, we find that only the Biaka,
Yoruba, Hakka, Cambodians, and Nasioi are not in strong LD (p ≤.04). Finally, across the third segment, spanning a distance of ∼7.4 kb, we find that only
the Biaka, Yoruba, Ibo, Hausa, African Americans, and Cambodians are not in strong and statistically significant LD (p ≤.03).

Blocks of Disequilibrium

In recent years there has been a good deal of discussion in the literature about the generality of blocks of LD (Jeffreys et al,35; Reich et al,50). A block is a
region of strong LD among multiple polymorphic sites such that in the population only a very few haplotypes (e.g. 2 or 3 of the many possible
haplotypes) account for the majority of the chromosomes. For the 7 sites we have studied at PAH, 128 haplotypes are possible. In our global survey we
have inferred that 93 occur somewhere in the world, and that 19 of those occur at a frequency of >7% in at least 1 population. In Europeans, only 4 of
these haplotypes are required to account for >75% of the chromosomes, and only 2 (BglII or PvuIIa plus MspI or XmnI) of the 7 sites are sufficient to
identify and discriminate among these 4 haplotypes. In this sense, this segment qualifies as a block. However, the issue of generality is not addressed
by the findings in European populations; the global situation is much more complex. In most African populations it takes many haplotypes to account
for even 70% of the chromosomes. In African Americans, 9 different haplotypes are required to account for 70% of the chromosomes and 5 of the 7
sites are required to identify and discriminate among them: the EcoRI site can be omitted and either the MspI or the XmnI site can be omitted. At the
other end of the spectrum, in East Asia, it takes only 2 or 3 haplotypes to account for 75% of the chromosomes depending on the specific population
and merely 2 sites [either BglII or PvuII(a) plus AluI] need be typed to discriminate between those haplotypes. This finding in East Asia would suit the
argument for typing "only a few sites," but the whole region is relatively uninformative. In North America, 3­5 haplotypes account for 75% of the
chromosomes and it would require typing 3 sites [either BglII or PvuII(a) plus PvuII(b) plus either MspI or XmnI]. Finally, in South America, 6­9
haplotypes account for 75% of the chromosomes in our sample, and it would require typing at least 4 sites [BglII, PvuII(a), and PvuII(b) plus either MspI
or XmnI]. Thus, we see that in some populations (from Africa), discrimination at only the 75% level requires 5 sites; in others (from South America), 4
sites are required; and in others (from Europe and East Asia), 2 sites are required, but not the same 2. From this we cannot conclude that these data
support or do not support a block structure because we have not identified clear limits to a block. What we can say based on these data is that blocks
and the SNPs needed to distinguish among haplotypes are not universal. However, when strong LD does exist among markers in a chromosomal
region in a population, it is possible to identify particular sites that capture the majority of the variation in that population at the haplotype level
whether or not there is a clearly defined single block. When we look at Native Americans in whom LD is even stronger, we see that the sites necessary
for discrimination are different from those in Europeans.

At this locus, "only a few sites" applies even if we are examining two separate blocks in Europeans. Finally, more uniform coverage (i.e., additional
markers in the large central intervals) will be necessary to determine whether the decline in LD across the locus is evenly correlated with distance or
confined to a small segment. If LD breaks down abruptly anywhere in the gene, it can only be somewhere in the PvuIIb to EcoRI 34.0­kb segment where
we have not yet examined closely spaced (<10 kb apart) markers.

It interesting to note that the haplotype frequencies of African Americans and of European Americans are very similar to those of the African regional
average for seven African populations and to the European regional average for six European populations. We propose that there are at least two
Downloaded 2022­2­28 8:15 P Your IP is 189.6.248.104
possible explanations:
The Population (1) of
Genetics thePAH,
African Americans
Judith R. Kidd;and the European
Kenneth K. KiddAmericans are compound mixtures of several populations from their continents of
Page 13 / 21
origin,
©2022and/or (2)Hill.
McGraw the likely substructuring
All Rights Reserved. that exists
Terms of among Americans
Use • Privacy (African
Policy Americans
• Notice and European Americans) is not represented in our study
• Accessibility
because of our sampling strategy. These explanations are closely related, and we think that both are true. The African Americans are from the Coriell
collection, and the European Americans have been sampled from a wide geographic and ethnic (within European) ancestry. Obviously, if we had
markers in the large central intervals) will be necessary to determine whether the decline in LD across the locus is evenly correlated with distance or
Sociedade Brasileira de Genetica Medica
confined to a small segment. If LD breaks down abruptly anywhere in the gene, it can only be somewhere in the PvuIIb to EcoRI 34.0­kb segment where
Access Provided by:
we have not yet examined closely spaced (<10 kb apart) markers.

It interesting to note that the haplotype frequencies of African Americans and of European Americans are very similar to those of the African regional
average for seven African populations and to the European regional average for six European populations. We propose that there are at least two
possible explanations: (1) the African Americans and the European Americans are compound mixtures of several populations from their continents of
origin, and/or (2) the likely substructuring that exists among Americans (African Americans and European Americans) is not represented in our study
because of our sampling strategy. These explanations are closely related, and we think that both are true. The African Americans are from the Coriell
collection, and the European Americans have been sampled from a wide geographic and ethnic (within European) ancestry. Obviously, if we had
selected these samples from a specific community in the United States where the substructure of the population(s) is more intact, the frequencies
might track more closely those of a specific group in Africa or Europe.

PAHdb curates a listing of haplotypes observed at the PAH locus. For the 6 polymorphisms in common, we have observed or inferred in the
populations we have studied all but 1 of the haplotypes in that listing. We have not observed haplotype 30 (46), 221112. In addition to nearly all of the
haplotypes in PAHdb, we have also observed or inferred 24 haplotypes not included in PAHdb (Table 101­4). None of these unlisted haplotypes occurs
in our data at a high frequency except, perhaps, 122212, which we have observed or inferred to be present in Japanese at 8.4% and in Yemenite Jews at
4.6%. Of these "novel" haplotypes, 8 were observed or inferred in only 1 population. The remainder were observed or inferred in 2­6 populations.
Table 101­4 shows these novel haplotypes, the populations in which they were observed or inferred, and their frequency in those populations.

PAH haplotype nomenclature has historically been a number assigned for each combination of alleles at the set of RFLPs detected using the cDNA as a
probe. These include sites we have not yet studied: HindIII (VNTR) and EcoRV downstream of the 3′ end of the gene. While that nomenclature is simple,
it will not survive as additional polymorphisms are studied in different combinations. In fact, most polymorphisms across the gene are not restriction
site polymorphisms and these more recently discovered polymorphisms may predominate in future studies of PAH. Defining haplotypes in these
future studies will require specificity about which polymorphisms are included in a study and using strings of allelic symbols to name the individual
haplotypes. Clearly, the summaries presented here for seven sites will be superseded by more thorough analyses of larger numbers of SNPs as our
studies of human genetic variation become ever more refined.

Implications for Haplotype Evolution

Some of the history of how mutations accumulated is irretrievable because the original mutations are old; the sites are polymorphic throughout Africa
and obviously predate the expansion of modern humans out of Africa. Sufficient recombination and drift have occurred that one cannot be sure of the
sequence of mutation and recombination events that generated the existing set of common haplotypes. Still, some speculation is possible if one
assumes that most of the common haplotypes have not involved recombination. One such possible scenario is outlined in Fig. 101­13. This explains
many of the common haplotypes but not the 1212212 (aqua) haplotype, which could be generated by a crossover between the common 1211121
(orange) and 2112212 (pale yellow) haplotypes. The much lower levels of LD across the 30­kb segment between the PvuIIb and EcoRI sites suggests that
this is a region of significant recurrent recombination in older populations.

Fig. 101­13

A possible evolutionary scheme for the origin of many common PAH haplotypes. The arrows in the main schema indicate accumulation of mutations as
indicated, using initial letters to specify the sites. Some common haplotypes cannot be explained by simple accumulation of mutations and require
either a crossover or recurrent mutation. Recombination seems more likely and one such example is indicated in the upper left. The color coding
corresponds to the colors used for these haplotypes in Figures 2 through 7.

Downloaded 2022­2­28 8:15 P Your IP is 189.6.248.104

The Population Genetics of PAH, Judith R. Kidd; Kenneth K. Kidd Page 14 / 21
©2022 McGraw Hill. All Rights Reserved. Terms of Use • Privacy Policy • Notice • Accessibility

Implications for Diversification of Modern Humans

A possible evolutionary scheme for the origin of many common PAH haplotypes. The arrows in the main schema indicate accumulation of mutations as
Sociedade Brasileira de Genetica Medica
indicated, using initial letters to specify the sites. Some common haplotypes cannot be explained by simple accumulation of mutations and require
Access Provided by:
either a crossover or recurrent mutation. Recombination seems more likely and one such example is indicated in the upper left. The color coding
corresponds to the colors used for these haplotypes in Figures 2 through 7.

Implications for Diversification of Modern Humans

Although PAH is a single locus, and human diversification cannot be entirely explicated by any single locus, data presented in this study support the
following:

CONCLUSIONS/FUTURE STUDIES
PAH is of interest to population genetics because of its multiple polymorphisms and pattern of geographic variation in allele and haplotype
frequencies that illuminate aspects of human evolution. Several interesting questions remain indeed, they are made more interesting by the results
presented here. For example, is there a position in the middle of the gene where LD breaks down across a very short distance? There are now more
SNPs known (see dbSNP, at http://www.ncbi.nlm.nih.gov/SNP/index.html for example) and these could now be tested to provide much denser
coverage of the gene. With respect to human evolution, questions arise about the locations of the separate founder events evident in the Native
American and East Asian populations. Will more populations from Siberia and Central Asia clarify these aspects of human diversification? These data
should also help clinical researchers identify the likely haplotypes containing new PKU mutations and should allow studies to determine whether
certain haplotypes seem more prone to deleterious mutation.

The understanding of clinically relevant variation can benefit from description of the normal variation at the locus. Furthermore, if there are common
functionally subtle molecular variants, they may well show LD with already known variants. Soon we will be able to track the evolution of the segments
of the gene i.e., which regions of which chromosomes share identity by descent among populations from different parts of the world.

Thus, years after this locus was the basis for some of the earliest molecular haplotype studies, new data show that there are still valuable lessons that
can be learned from studies of this important locus.

ACKNOWLEDGMENTS
This work was funded in part by National Institutes of Health grants GM57672 and AA09379 to KKK and NSF BCS­9912028 to JRK. We thank Andrew J.
Pakstis, William C. Speed, Roy Capper, and Valeria Ruggeri for their excellent technical assistance. We are indebted to the following people who helped
assemble the diverse population collection used in this study: F.L. Black, L.L. Cavalli­Sforza, Batsheva Bonne­Tamir, K. Dumars, J. Friedlaender, David
Goldman, K. Kendler, W. Knowler, F. Oronsaye, J. Parnas, L. Peltonen, L.O. Schulz and K. Weiss, Elena Grigorenko, Sylvester L.B. Kajuna, Nganyirwa J.
Karoma, Selemani Kungulilo, Ru­Band Lu, Kunle Odunsi, Friday Okonofua, and Olga V. Zhukova.

In addition, some of the cell lines were obtained from the National Laboratory for the Genetics of Israeli Populations at Tel Aviv University, Israel, and
the African­American samples were obtained from the Coriell Institute for Medical Research, Camden, NJ. Special thanks are due to the many
hundreds of individuals who volunteered to give blood samples for studies such as this. Without such participation of individuals from diverse parts of
the world we would be unable to obtain a good picture of the genetic variation in our species.

Downloaded 2022­2­28 8:15 P Your IP is 189.6.248.104

REFERENCES
The Population Genetics of PAH, Judith R. Kidd; Kenneth K. Kidd Page 15 / 21
©2022 McGraw Hill. All Rights Reserved. Terms of Use • Privacy Policy • Notice • Accessibility
1. Anderson MA, Gusella JF: Use of cyclosporin A in establishing Epstein­Barr virus­transformed human lymphoblastoid cell lines. In Vitro. 1984;
20:856.
In addition, some of the cell lines were obtained from the National Laboratory for the Genetics of Israeli Populations
Sociedadeat Tel Aviv University,
Brasileira Israel,Medica
de Genetica and
the African­American samples were obtained from the Coriell Institute for Medical Research, Camden, NJ. Special thanks
Access Provided by:are due to the many
hundreds of individuals who volunteered to give blood samples for studies such as this. Without such participation of individuals from diverse parts of
the world we would be unable to obtain a good picture of the genetic variation in our species.

REFERENCES

1. Anderson MA, Gusella JF: Use of cyclosporin A in establishing Epstein­Barr virus­transformed human lymphoblastoid cell lines. In Vitro. 1984;
20:856.
[PubMed: 6519667]

2. Apold J, Eiken HG, Odland E, Fredriksen A, Bakken A, Lorens JB, Boman H: A termination mutation prevalent in Norwegian haplotype 7
phenylketonuria genes. Am J Hum Genet. 1990; 47:1002.
[PubMed: 1978553]

3. Avigad S, Cohen BE, Bauer S, Schwartz G, Frydman M, Woo SL, Niny Y, Shiloh Y: A single origin of phenylketonuria in Yemenite Jews. Nature.
1990; 344:168.
[PubMed: 1968617]

4. Baric I, Mardesic D, Gjuric G, Sarnavka V, Gobel­Schreiner B, Lichter­Konecki U, Konecki DS, et al.: Haplotype distribution and mutations at the
PAH locus in Croatia. Hum Genet. 1992; 90:155.
[PubMed: 1358784]

5. Bickel H, Bachmann C, Beckers R: Neonatal mass screening for metabolic disorders: A collaborative study. Eur J Pediatr. 1981; 137:133.

6. Capon F, Munro M, Barker J, Trembath R: Searching for the major histocompatibility complex psoriasis susceptibility gene. J Invest Dermatol.
2002; 118:745.
[PubMed: 11982750]

7. Ceppellini RM, Siniscalco M, Smith CAB: The estimation of gene frequencies in a random­mating population. Ann Hum Genet. 1955; 20:97.
[PubMed: 13268982]

8. Chakraborty R, Lidsky AS, Daiger SP, Guttler F, Sullivan S, DiLella AG, Woo SLC: Polymorphic DNA haplotypes at the human phenylalanine
hydroxylase locus and their relationship with phenylketonuria. Hum Genet. 1987; 76:40.
[PubMed: 2883110]

9. Chattopadhyay P, Pakstis AJ, Mukherjee N, Iyengar S, Odunsi A, Okonofua F, Bonne­Tamir B, Speed W, Kidd JR, Kidd KK: A global survey of
haplotype frequencies and linkage disequilibrium at the RET locus. Eur J Hum Genet , in press.

10. Cheung KH, Nadkarni P, Silverstein S, Kidd JR, Pakstis AJ, Miller P, Kidd KK: PhenoDB: An integrated client/server database for linkage and
population genetics. Comput Biomed Res. 1996; 29:327.
[PubMed: 8812078]

11. Daiger SP, Chakraborty R, Reed L, Fekete G, Schuler D, Berenssi G, Nasz I, et al.: Polymorphic DNA haplotypes at the phenylalanine hydroxylase
(PAH) locus in European families with phenylketonuria (PKU). Am J Hum Genet. 1989a; 45:310.
[PubMed: 2569271]

12. Daiger SP, Reed L, Huang S­S, Zeng Y­T, Wang T, Lo WHY, Okano Y, et al.: Polymorphic DNA haplotypes at the phenylalanine hydroxylase (PAH)
locus in Asian families with phenylketonuria (PKU). Am J Hum Genet. 1989b; 45:319.
[PubMed: 2569272]

13. Degioanni A, Darlu P: Analysis of the molecular variance at the phenylalanine hydroxylase (PAH) locus. Eur J Hum Genet. 1994; 2:166.
[PubMed: 7834276]

14. DeLella AG, Marvit J, Brayton K, Woo SL: An amino acid substitution involved in phenylketonuria is in linkage disequilibrium with DNA haplotype
2. Nature. 1987;
Downloaded 327:333.8:15 P Your IP is 189.6.248.104
2022­2­28
[PubMed:
The 2884570]
Population Genetics of PAH, Judith R. Kidd; Kenneth K. Kidd Page 16 / 21
©2022 McGraw Hill. All Rights Reserved. Terms of Use • Privacy Policy • Notice • Accessibility
15. DeLella AG, Marvit J, Lidsky AS, Guttler F, Woo SL: Tight linkage between a splicing mutation and a specific DNA haplotype in phenylketonuria.
Nature. 1986; 322:799.
 Sociedade
13. Degioanni A, Darlu P: Analysis of the molecular variance at the phenylalanine hydroxylase (PAH) locus. Eur Brasileira
J Hum Genet. 1994;de Genetica Medica
2:166.
Access Provided by:
[PubMed: 7834276]

14. DeLella AG, Marvit J, Brayton K, Woo SL: An amino acid substitution involved in phenylketonuria is in linkage disequilibrium with DNA haplotype
2. Nature. 1987; 327:333.
[PubMed: 2884570]

15. DeLella AG, Marvit J, Lidsky AS, Guttler F, Woo SL: Tight linkage between a splicing mutation and a specific DNA haplotype in phenylketonuria.
Nature. 1986; 322:799.
[PubMed: 3018584]

16. DeMille MMC, Kidd JR, Ruggeri V, Palmatier MA, Goldman D, Odunsi A, Okonofua F, Grigorenko E, Schulz LO, Bonne­Tamir B, Lu RB, Parnas J,
Pakstis AJ, Kidd KK: Population variation in linkage disequilibrium across the COMT gene considering promoter region and coding region variation.
Hum Genet. 2002; 111:521.
[PubMed: 12436243]

17. Dempster AP, Laird NM, Rubin DB: Maximum likelihood from incomplete data via the EM algorithm. J R Stat Soc B. 1977; 39:1.

18. Devlin B, Risch N: A comparison of linkage disequilibrium measures for fine­scale mapping. Genomics. 1995; 29:311.
[PubMed: 8666377]

19. Dianzani I, Devoto M, Camaschella C, Saglio G, Ferrero GB, Cerone R, Romano C, et al.: Haplotype distribution and molecular defects at the
phenylalanine hydroxylase locus in Italy. Hum Genet. 1990; 86:69.
[PubMed: 1979309]

20. Dworniczak B, Wedemeyer N, Eiel A, Horst J: PCR detection of the PvuII (Ea) RFLP at the human phenylalanine hydroxylase (PAH) locus. Nucleic
Acids Res. 1991a; 19:1958.
[PubMed: 1674373]

21. Dworniczak B, Wedemeyer N, Horst J: PCR detection of the BglII RFLP at the human phenylalanine hydroxylase (PAH) locus. Nucleic Acids Res.
1991b; 19:1958.
[PubMed: 1674372]

22. Eisensmith RC, Okano Y, Dasovich M, Wang T, Guttler F, Lou H, Guldberg P, et al.: Multiple origins for phenylketonuria in Europe. Am J Hum
Genet. 1992; 51:1355.
[PubMed: 1361100]

23. Fallin D, Shork NJ: Accuracy of haplotype frequency estimation for biallelic loci, via the expectation­maximization algorithm for unphased diploid
genotype data. Am J Hum Genet. 2000; 67:947.
[PubMed: 10954684]

24. Feingold J, Guilloud­Bataille M, Feingold N, Rey F, Berthelon M, Lyonnet S: Linkage disequilibrium in the human phenylalanine hydroxylase
locus. Dev Brain Dysfunct. 1993; 6:26.

25. Frisse L, Hudson RR, Bartozewicz A, Wall JD, Donfack J, DiRienzo A: Gene conversion and different population histories may explain the contrast
between polymorphism and linkage disequilibrium levels. Am J Hum Genet. 2001; 69:831.
[PubMed: 11533915]

26. Gabriel SB, Schaffner SF, Nguyen H, Moore JM, Roy J, Blumenstiel B, Higgins J, DeFelice M, Lochner A, Faggart M, Liu­Cordero SN, Rotimi C,
Adeyemo A, Cooper R, Ward R, Lander ES, Daly MJ, Altshuler D: The structure of haplotype blocks in the human genome. Science. 2002; 296:2225.
[PubMed: 12029063]

27. Goltsov AA, Eisensmith RC, Naughton ER, Jin L, Chakraborty R, Woo SLC: A single polymorphic STR system in the human phenylalanine
hydroxylase gene permits rapid prenatal diagnosis and carrier screening for phenylketonuria. Hum Mol Genet. 1993; 2:577.
[PubMed: 8100164]
Downloaded 2022­2­28 8:15 P Your IP is 189.6.248.104
28.
TheGoltsov AA, Eisensmith
Population Genetics ofRC, Woo
PAH, SLC: R.
Judith Detection of the XmnI
Kidd; Kenneth RFLP at the human PAH locus by PCR. Nucleic Acids Res. 1992; 20:927.
K. Kidd Page 17 / 21
[PubMed:
©2022 1347420]
McGraw Hill. All Rights Reserved. Terms of Use • Privacy Policy • Notice • Accessibility

29. Hawley ME, Kidd KK: HAPLO: A program using the EM algorithm to estimate the frequencies of multi­site haplotypes. J Hered. 1995; 86:409.
[PubMed: 12029063]
Sociedade Brasileira de Genetica Medica
27. Goltsov AA, Eisensmith RC, Naughton ER, Jin L, Chakraborty R, Woo SLC: A single polymorphic STR system
Access in the human
Provided by: phenylalanine
hydroxylase gene permits rapid prenatal diagnosis and carrier screening for phenylketonuria. Hum Mol Genet. 1993; 2:577.
[PubMed: 8100164]

28. Goltsov AA, Eisensmith RC, Woo SLC: Detection of the XmnI RFLP at the human PAH locus by PCR. Nucleic Acids Res. 1992; 20:927.
[PubMed: 1347420]

29. Hawley ME, Kidd KK: HAPLO: A program using the EM algorithm to estimate the frequencies of multi­site haplotypes. J Hered. 1995; 86:409.
[PubMed: 7560877]

30. Hertzberg M, Jahromi K, Ferguson V, Dahl HHM, Mercer J, Mickleson KNP, Trent RJ: Phenylalanine hydroxylase gene haplotypes in Polynesians:
Evolutionary origins and absence of alleles associated with severe phenylketonuria. Am J Hum Genet. 1989; 44:382.
[PubMed: 2563633]

31. Hofman KJ, Steel G, Kazazian HH, Vallie D: Phenylketonuria in U.S. blacks: Molecular analysis of the phenylalanine hydroxylase gene. Am J Hum
Genet. 1991; 48:791.
[PubMed: 2014802]

32. Horikawa Y, Oda N, Cox NJ, Li X, Orho­Melander M, Hara M, Hinokio Y, Lindner TH, Mashima H, Schwarz PE, del Bosque­Plata L, Oda Y,
Yoshiuchi I, Colilla S, Polonsky KS, Wei S, Concannon P, Iwasaki N, Schulze J, Baier LJ, Bogardus C, Groop L, Boerwinkle E, Hanis CL, Bell GI:
Genetic variation in the gene encoding calpain­10 is associated with type 2 diabetes mellitus. Nat Genet. 2000; 26:163.
[PubMed: 11017071]

33. Iyengar S, Seaman M, Deinard AS, Rosenbaum HC, Sirugo G, Castiglione CM, Kidd JR, et al.: Analyses of cross­species polymerase chain reaction
products to infer the ancestral sate of human polymorphisms. DNA Seq. 1998; 8:317.
[PubMed: 10993602]

34. Jaruzelska J, Henriksen K, Guttler F, Riess O, Borski K, Blin N, Slomski R: The codon 408 mutation associated with haplotype 2 is predominant in
Polish families with phenylketonuria. Hum Genet. 1991; 86:247.
[PubMed: 1671768]

35. Jeffreys AJ, Kauppi L, Neuman R: Intensely punctate meiotic recombination in the class II region of the major histocompatibility complex. Nat
Genet. 2001; 29:217.
[PubMed: 11586303]

36. Jorde LB, Watkins WS, Carlson M, Groden J, Albertsen H, Thuveris A, Leppert M: Linkage disequilibrium predicts physical distance in the
adenomatous polyposis coli region. Am J Hum Genet. 1994; 54:884.
[PubMed: 8178829]

37. Jorde LB, Watkins WS, Kere J, Nyman D, Eriksson AW: Gene mapping in isolated populations: New roles for old friends?? Hum Hered. 2000; 50:57.
[PubMed: 10545758]

38. Kidd JR, Pakstis AJ, Zhao H, Lu R­B, Okonofua FE, Odunsi A, Grigorenko E, et al.: Haplotypes and linkage disequilibrium at the phenylalanine
hydroxylase locus, PAH, in a global representation of populations. Am J Hum Genet. 2000; 66:1882.
[PubMed: 10788337]

39. Kidd KK: Phenylketonuria: The population genetics of a disease. Nature. 1987; 327:282.
[PubMed: 2884567]

40. Kidd KK, Morar B, Castiglione CM, Zhao H, Pakstis AJ, Speed WC, Bonne­Tamir B, Lu RB, Goldman D, Lee C, Nam YS, Grandy DK, Jenkins T,
Kidd JR: A global survey of haplotype frequencies and linkage disequilibrium at the DRD2 locus Hum Genet. 1998; 103:211.
[PubMed: 9760208]

41. Kim SJ, Cox N, Courchesne R, Lord C, Corsello C, Akshoomoff N, Guter S, Leventhal BL, Courchesne E, Cook Jr EH: Transmission disequilibrium
mapping at the serotonin transporter gene (SLC6A4) region in autistic disorder. Mol Psychiatry. 2002; 7:278.
Downloaded 2022­2­28 8:15 P Your IP is 189.6.248.104
[PubMed:
The 11920155]
Population Genetics of PAH, Judith R. Kidd; Kenneth K. Kidd Page 18 / 21
©2022 McGraw Hill. All Rights Reserved. Terms of Use • Privacy Policy • Notice • Accessibility
42. Kittles RA, Young D, Weinrich S, Hudson J, Argyropoulos G, Ukoli F, Adams­Campbell L, Dunston GM: Extent of linkage disequilibrium between
the androgen receptor gene CAG and GGC repeats in human populations: Implications for prostate cancer risk. Hum Genet. 2001; 109:253.
40. Kidd KK, Morar B, Castiglione CM, Zhao H, Pakstis AJ, Speed WC, Bonne­Tamir B, Lu RB, Goldman D, Lee C, Nam YS, Grandy DK, Jenkins T,
Kidd JR: A global survey of haplotype frequencies and linkage disequilibrium at the DRD2 locus Hum Genet.Sociedade
1998; 103:211.Brasileira de Genetica Medica
Access Provided by:
[PubMed: 9760208]

41. Kim SJ, Cox N, Courchesne R, Lord C, Corsello C, Akshoomoff N, Guter S, Leventhal BL, Courchesne E, Cook Jr EH: Transmission disequilibrium
mapping at the serotonin transporter gene (SLC6A4) region in autistic disorder. Mol Psychiatry. 2002; 7:278.
[PubMed: 11920155]

42. Kittles RA, Young D, Weinrich S, Hudson J, Argyropoulos G, Ukoli F, Adams­Campbell L, Dunston GM: Extent of linkage disequilibrium between
the androgen receptor gene CAG and GGC repeats in human populations: Implications for prostate cancer risk. Hum Genet. 2001; 109:253.
[PubMed: 11702204]

43. Konecki DS, Lichter­Konecki U: The phenylketonuria locus: current knowledge about alleles and mutations of the phenylalanine hydroxylase gene
in various populations. Hum Genet. 1991; 87:377.
[PubMed: 1679029]

44. Kozak L, Kuhrova V, Blazkova M, Tomano V, Fajkusova L, Dvorakova D, Pijackova A: Phenylketonuria mutations and their relation to RFLP
haplotypes at the PAH locus in Czech PKU families. Hum Genet. 1995; 96:472.
[PubMed: 7557973]

45. Lidsky AS, Ledley FD, DiLella AG, Kwok SCM, Daiger SP, Robson KJH, Woo SLC: Extensive restriction site polymorphism at the human
phenylalanine hydroxylase locus and application in prenatal diagnosis of phenylketonuria. Am J Hum Genet. 1985; 37:619.
[PubMed: 9556654]

46. Lonjou C, Collins A, Ajioka RS, Jorde LB, Kushner JP, Morton NE: Allelic association under map error and recombinational heterogeneity: A tale
of two sites. Proc Natl Acad Sci USA. 1999; 95:11366.
[PubMed: 9736742]

47. Myers A, Holmans P, Marshall H, Kwon J, Meyer D, Ramic D, Shears S, Booth J, DeVrieze FW, Crook R, Hamshere M, Abraham R, Tunstall N, Rice
F, Carty S, Lillystone S, Kehoe P, Rudrasingham V, Jones L, Lovestone S, Perez­Tur J, Williams J, Owen MJ, Hardy J, Goate AM: Susceptibility locus
for Alzheimer’s disease on chromosome 10. Science. 2000; 290:2304.
[PubMed: 11125144]

48. Nowacki PM, Bick S, Prevost L, Scriver CR: The PAH mutation analysis consortium database: Update 1996. Nucleic Acids Res. 1997; 25:139.
[PubMed: 9016524]

49. Purandare SM, Cawthon R, Nelson LM, Sawada S, Watkins WS, Ward K, Jorde LB, Viskochil DH: Genotyping of PCR­based polymorphisms and
linkage disequilibrium analysis at the NF1 locus. Am J Hum Genet. 1996; 59:159.
[PubMed: 8659521]

50. Reich DE, Cargill M, Bolk S, Ireland J, Sabeti PC, Richter DJ, Lavery T, Kouyoumjian R, Farhadian SF, Ward R, Lander ES: Linkage disequilibrium
in the human genome. Nature. 2001; 411:199.
[PubMed: 11346797]

51. Sambrook J, Fritsch E, Maniatis T: Molecular Cloning: A Laboratory Manual, 2nd ed.Cold Spring Harbor , New York, Cold Spring Harbor Laboratory
Press, 1989.

52. Sawcer S, Maranian M, Setakis E, Curwen V, Akesson E, Hensiek A, Coraddu F, Roxburgh R, Sawcer D, Gray J, Deans J, Goodfellow PN, Walter
N, Clayton D, Compston A: A whole genome screen for linkage disequilibrium in multiple sclerosis confirms disease associations with regions
previously linked to susceptibility. Brain. 2002; 125:1337.
[PubMed: 12023322]

53. Scapoli L, Martinelli M, Pezzetti F, Carinci F, Bodo M, Tognon M, Carinci P: Linkage disequilibrium between GABRB3 gene and nonsyndromic
familial cleft lip with or without cleft palate. Hum Genet. 2002; 110:15.
[PubMed: 11810291]

54. Scriver CR,2022­2­28

Downloaded Byck S, Prevost
8:15 PL,Your
Hoang IP L,
is PAH Mutation Analysis Consortium: The phenylalanine hydroxylase locus: A marker for the history of
189.6.248.104
phenylketonuria
The and human
Population Genetics diversity,
of PAH, Judithin Chadwick D, Cardew
R. Kidd; Kenneth G (eds): Variation in the Human Genome. Ciba Foundation Symposium 1997.
K. Kidd Chicester,
Page 19 / 21
©2022
England,McGraw Hill. All
Wiley, 1996, ppRights
73–96. Reserved. Terms of Use • Privacy Policy • Notice • Accessibility

55. Stuhrmann M, Riess O, Monch E, Kurdoglu G: Haplotype analysis of the phenylalanine hydroxylase gene in Turkish phenylketonuria families. Clin
 Sociedade
53. Scapoli L, Martinelli M, Pezzetti F, Carinci F, Bodo M, Tognon M, Carinci P: Linkage disequilibrium between GABRB3 Brasileira
gene andde Genetica Medica
nonsyndromic
Access Provided by:
familial cleft lip with or without cleft palate. Hum Genet. 2002; 110:15.
[PubMed: 11810291]

54. Scriver CR, Byck S, Prevost L, Hoang L, PAH Mutation Analysis Consortium: The phenylalanine hydroxylase locus: A marker for the history of
phenylketonuria and human diversity, in Chadwick D, Cardew G (eds): Variation in the Human Genome. Ciba Foundation Symposium 1997. Chicester,
England, Wiley, 1996, pp 73–96.

55. Stuhrmann M, Riess O, Monch E, Kurdoglu G: Haplotype analysis of the phenylalanine hydroxylase gene in Turkish phenylketonuria families. Clin
Genet. 1989; 36:117.
[PubMed: 2569949]

56. Svensson E, Von Dobeln U, Hagenfeldt L: Polymorphic DNA haplotypes at the phenylalanine hydroxylase locus and their relation to phenotype in
Swedish phenylketonuria families. Hum Genet. 1991; 87:11.
[PubMed: 1674714]

57. Tishkoff SA, Dietzsch E, Speed W, Pakstis AJ, Kidd JR, Cheung K, Bonne­Tamir B, Santachiara­Benerecetti AS, Moral P, Krings M, et al.: Global
patterns of linkage disequilibrium at the CD4 locus and modern human origins. Science. 1996; 271:1380.
[PubMed: 8596909]

58. Tishkoff SA, Goldman A, Calafell F, Speed WC, Deinhard AS, Bonne­Tamir B, Kidd JR, Pakstis AJ, Jenkins T, Kidd KK: A global haplotype analysis
of the myotonic dystrophy locus: Implications for the evolution of modern humans and for the origin of myotonic dystrophy mutations. Am J Hum
Genet. 1998; 62:1389.
[PubMed: 9585589]

59. Tiskoff SA, Pakstis AJ, Ruano G, Kidd KK: The accuracy of statistical methods for estimation of haplotype frequencies: An example from the CD4
locus. Am J Hum Genet. 2000; 67:518.
[PubMed: 10859209]

60. Vaessen N, Heutink P, Houwing­Duistermaat JJ, Snijders PJ, Rademaker T, Testers L, Batstra MR, Sandkuijl LA, van Duijn CM, Oostra BA: A
genome­wide search for linkage­disequilibrium with type 1 diabetes in a recent genetically isolated population from the Netherlands. Diabetes. 2002;
51:856.
[PubMed: 11872691]

61. Wedemeyer N, Dworniczak B, Horst J: PCR detection of the MspI (Aa) RFLP at the human phenylalanine hydroxylase (PAH) locus. Nucleic Acids Res.
1991; 19:1959.
[PubMed: 1709499]

62. Woo SLC: Collation of RFLP haplotypes at the human phenylalanine hydroxylase (PAH) locus. Am J Hum Genet. 1988; 43:781.
[PubMed: 2903669]

63. Woo SL, Lidsky AS, Guttler F, Chandra T, Robson KJ: Cloned human phenylalanine hydroxylase gene allows prenatal diagnosis and carrier
detection of classical phenylketonuria. Nature. 1983; 306:151.
[PubMed: 6316140]

64. Zhang S, Pakstis AJ, Kidd KK, Zhao H: Comparisons of two methods for haplotype reconstruction and haplotype frequency estimation from
population data. Am J Hum Genet. 2001; 69:906.
[PubMed: 11536083]

65. Zhao H­Y, Pakstis AJ, Kidd JR, Kidd KK: Assessing linkage disequilibrium in a complex genetic system. I. Overall deviation from random
association. Ann Hum Genet. 1999; 63:167.
[PubMed: 10738528]

66. Zygulska M, Eigel A, Aulehla­Scholz C, Pietrzyk JJ, Horst J: Molecular analysis of PKU haplotypes in the population of southern Poland. Hum
Genet. 1991; 86:292.
[PubMed: 1671770]
Downloaded 2022­2­28 8:15 P Your IP is 189.6.248.104
The Population Genetics of PAH, Judith R. Kidd; Kenneth K. Kidd Page 20 / 21
©2022 McGraw Hill. All Rights Reserved. Terms of Use • Privacy Policy • Notice • Accessibility
association. Ann Hum Genet. 1999; 63:167.
Sociedade Brasileira de Genetica Medica
[PubMed: 10738528]
Access Provided by:

66. Zygulska M, Eigel A, Aulehla­Scholz C, Pietrzyk JJ, Horst J: Molecular analysis of PKU haplotypes in the population of southern Poland. Hum
Genet. 1991; 86:292.
[PubMed: 1671770]

Downloaded 2022­2­28 8:15 P Your IP is 189.6.248.104

The Population Genetics of PAH, Judith R. Kidd; Kenneth K. Kidd Page 21 / 21
©2022 McGraw Hill. All Rights Reserved. Terms of Use • Privacy Policy • Notice • Accessibility