GBIF metagenomics and metacrap

Yes, this is a clickbait headline, and yes, it may seem like shooting fish in a barrel to complain about crappy data in GBIF, but my point here is raise concerns about the impact of metagenomic data on GBIF, and how difficult it may be to track down the causes of errors.

I stumbled across this example while looking for specimen records for the genus Rafflesia, which are parasitic plants famous for the spectacular size of their flowers (up to 1m across).

The GBIF map for Rafflesia shows a few outliers. Unfortunately GBIF doesn't make it easy to drill down (why oh why can't we just click on the map and see the corresponding occurrences?) so I opened the map in iSpecies and clicked on each outlier in turn. The one in Vanuatu (438164267 from the Paris museum P00577336) is identified to genus level only and has the note:

Parasite terrestre, grande fleur orange au ras du sol. Incomplète suite à prédation. Très forte odeur désagréable. Récolté par Sylvain Hugel (photo) (alcool seul)

which Google translates as:

Terrestrial parasite, large orange flower at ground level. Incomplete due to predation. Very strong unpleasant odor. Collected by Sylvain Hugel (photo) (alcohol only)

Sounds a bit like Rafflesia but there's no photo or other information available online. Likewise there's no additional data for the record from Brazil (1090499968). There is a record from Madagascar that is accompanied by a photo, but it that looks nothing like Rafflesia (1261055923):

That leaves two records, 2018528337 (Rafflesia cantleyi) from off the coast of Peru, and 2014813273 (Rafflesia) off the coast of Australia. Both of these records are metagenomic. For example, occurrence 2018528337 is part of a dataset Amplicon sequencing of Tara Oceans DNA samples corresponding to size fractions for protists that, on the face of it, would be an unlikely source of occurrences of forest dwelling plants.

What we get in the GBIF occurrence record is a link to the pipeline used to generate the data (Pipeline version 4.1 - 17-Jan-2018), the sample (ERS491947), and an analysis (MGYA00167469) that summarises all the taxonomic data from the ocean water sample.

What we don't get in GBIF is an obvious way to try and figure out why GBIF thinks that large flowers live in the ocean. I followed the links from MGYA00167469 and downloaded a bunch of files, some in familiar formats (FASTA), others in formats I'd not seen before (e.g., HDF5). From the mapseq file we have the following line:

ERR562574.2494029-BISMUTH-0000:2:112:3465:9749-2/89-1 GFBU01000011.4303.6106 76 0.9634146094322205 79 3 0 6 88 1722 1804 +  sk__Eukaryota;k__Viridiplantae;p__Streptophyta;c__;o__Malpighiales;f__Rafflesiaceae;g__Rafflesia;s__Rafflesia_cantleyi 

This tells us that sequence ERR562574.2494029-BISMUTH-0000:2:112:3465:9749-2/89-1 matches GenBank sequence GFBU01000011, which is "Rafflesia cantleyi RC_11 transcribed RNA sequence" from a paper on flower development in Rafflesia cantleyi doi:10.1371/journal.pone.0167958. So, now we see why we think we have giant flowers off the coast of Peru.

The rest of the line has information on the match: the oceanic sequence has a 0.96 identity with the plant sequence, has 79 matches, 3 mismatches, and no gaps, which suggests that this is a short sequence. Going digging in the FASTA file I found the raw sequence, and it is indeed very short:

>ERR562574.2494029-BISMUTH-0000:2:112:3465:9749-2/89-1
GTCTAAGTGTCGTGAGAAGTTCGTTGAACCTGATCATTTAGAGGAAGTAGAAGTCGTAAC
AAGGTTTCCGTAGGTGAACCTGCGGAAGG

This short string is the evidence for Rafflesia in the ocean. Out of curiosity I ran this sequence through BLAST:

Query  1     GTCTAAGTGTCGTGAGAAGTTCGTTGAACCTGATCATTTAGAGGAAGTAGAAGTCGTAAC  60
             |||||||||||||| |||||||||||||||| ||||||||||||||| ||||||||||||
Sbjct  1683  GTCTAAGTGTCGTGGGAAGTTCGTTGAACCTTATCATTTAGAGGAAGGAGAAGTCGTAAC  1742

Query  61    AAGGTTTCCGTAGGTGAACCTGCGGAAGG  89
             |||||||||||||||||||||||||||||
Sbjct  1743  AAGGTTTCCGTAGGTGAACCTGCGGAAGG  1771

The top hit is Bathycoccus prasinos, a picoplanktonic alga with a world-wide distribution. This seems like a more plausible identification for this sequence (all the top 100 hits are very similar to each other, many are labelled as Bathycoccus).

So, there's something clearly amiss with the analysis of this dataset. Someone who knows more about metagenomics than I do will be better placed to explain why this pipeline got it so wrong, and how common this issue is.

Given the scale and automation of metagenomics, there will always be errors - that is inevitable. What we need is ways to catch those errors, especially ones that are going to "pollute" existing distribution data with spurious records (flowers in the ocean). And in a sense, GBIF excels at this in that it exposes data to a wider audience. If you work on marine microbiology you might not notice that your sequences apparently include forest plants, but if you work on forest plants you will almost certainly notice sequences occurring in the ocean.

A key feature of GBIF that makes it so useful is that, unlike many data repositories, it does not treat data as a "black box". GBIF is not like a library catalogue which merely tells you that they have books and where to find them, instead it is like Google Books, which can tell you which books contain a given phrase you are looking for. By opening up each dataset and indexing the contents, GBIF enables us to do data analysis (in much the same way that GenBank isn't just a catalogue of sequences, it enables you to search the sequences themselves).

This is a feature we risk losing if we treat metagenomics data as a black box. The Tara Oceans data that GBIF receives is simply a list of taxa at a locality, it's a checklist. We have to take it on trust that the taxonomic assignments are accurate, and it is not a trivial task to diagnose errors. Compare this to having the photo that accompanied the record from Madagascar, which helps us determine that the identification is wrong. Going forward, it would be helpful if we had metagenomic sequences available as part of the data download from GBIF. It's also worth considering whether GBIF should start doing its own analysis of sequence data, or asking its contributors to check that their taxonomic assignments are correct (e.g., running BLAST on the data). Otherwise GBIF users may end up having to filter their data for a growing number of completely spurious records.

Update

Looks like (occurrence 1261055923) is Langsdorffia:

Related to the blog post https://t.co/do5P2u5Ooj, does anyone have ideas on the identity of this species (@GBIF record https://t.co/zSr8xemkh7 from Madagascar) which is labelled "Rafflesia" pic.twitter.com/qiRtYH0tUh

— Roderic Page (@rdmpage) December 20, 2019

Metacrap

Time for a rant. I spend a lot of time fussing with records from sources such as GenBank and DiGIR providers, trying to extract strings that might be identifiers, with a view to linking sequences to specimens (and thus to localities), sequences to publications, publications to GUIDs, etc. The goal is to be able to link all these resources together, so that I can go seamlessly from a phylogeny in TreeBASE to a locality on a map, and visa versa, of from a phylogeny for a parasite to that of its host, for example.

Time and again I come across extraordinarily frustrating messes. More and more I see why Cory Doctorow wrote Metacrap. According to this essay, the problems with metadata are:

1. People lie
2. People are lazy
3. People are stupid
4. Mission: Impossible -- know thyself
5. Schemas aren't neutral
6. Metrics influence results
7. There's more than one way to describe something

Amen. I know I've moaned about this before (e.g., Damn DiGIR, and AMNH Dspace and Openurl), but it's just amazing what people put into databases.

Co-ordinates

My current favourite is the junk that populates GenBank source records. This is where people provide information about the source of their sequence (e.g., the organism it came from, any voucher specimens, and the geographic locality of the source). All manner of junk ends up here.

You'd think that geographic location was a straightforward thing to record, but no. Consider sequence AY281248. The country field for this record is:

Australia: Gubbata, NSW (GPS: 33 38' 07'', 146  33' 12'')

Spot the problem? Leaving aside the fact that somebody has to parse this and extract the latitude and longitude (hmm, which one is which, can I rely on latitude coming first, ah the second must be longitude because latitude is never > 90°), the co-ordinates are not in Australia.


View Larger Map

Last time I checked, Australia was in the southern hemisphere, and not drifting of the coast of Japan. Now, latitude 33°38' 07''S is in Australia:


View Larger Map

Seems like a small deal to leave off "S" (or a minus sign), but changing hemispheres is a big deal. Then there are strings such as

Nicaragua: Rio San Juan, Near Isla de Diamante 
(ca. 15 km SE El Castillo on Rio San Juan), 10deg56'N 84deg18'W

from DQ502492. Sigh.

Now, you might say that this isn't GenBank's fault, because the country field wasn't supposed to have latitude and longitude information, that is the role of the lat_lon field. This format for this field is defined as:

degrees latitude and longitude in format "d[d.dd] N|S d[dd.dd] W|E"

Fine, but people don't always follow this format. For example, DQ226041 has

/lat_lon="6 28.06'N; 58 37.16'W"

Not hugely different, but the ";" needs to be removed before parsing. What's worse, I can't assume people will follow GenBank guidelines about how to store data in these fields.

Specimen codes

What I'm really after are specimen codes, as these can potentially be linked to digital records served by natural history collections. But, once again all manner of variations end up in GenBank records. For example, specimens in the Field Museum can be referred to as Field Museum of Natural History 167358 (DQ023431) or as FMNH 167358 ( AY324464 ), and there are other variations I've come across as well. Furthermore, FMNH 167358 isn't enough by itself to retrieve a record from the Field Museum, I need to know the address of the DiGIR provider, and whether the specimen is a mammal or some other vertebrate (museum specimen codes are often not unique across the institution). In my opinion the single biggest failure of specimen digitisation efforts is the lack of a globally unique, resolvable identifier. Sure, it's coming, but in the meantime we have chaos.

Tools like GBIF show us what can be done when we aggregate lots of specimen records, but I'd argue the real fun starts when we link resources together, and we can only do this is we have stable, shared identifiers. As an example, Steppan et al. (2003) (doi:10.1111/j.1095-8312.2003.00274.x) published a study of Apomys biogeography. If I get a sequence from this study from GenBank (e.g., AY324464 mentioned above) I have a GenBank record with very few links. There's no PubMed record, for example. If I use an OpenURL resolver (such as the one I wrote for bioGUID) I can retrieve a DOI from the citation Biol. J. Linn. Soc. Lond. 80 (4), 699-715 (2003). This gives me a GUID for the paper, as well as a way to link to the text. The specimen code FMNH 167358 can be used to retrieve details from the Field Museum (once I've parsed the GenBank taxonomy string to discover that Apomys datae is a mammal). This gives me the latitude and longitude of the locality from which the specimen was collected, so I can put it on a map:


View Larger Map

So, I've added information to the GenBank record. But this isn't really the goal. What I want to do is link these elements together. For example, here's a graph showing the link between the publication, the sequence, and the specimen:



Basically, this is the immediate neighbourhood of the sequence AY324464. If I get the neighbourhood of the specimen:



Note that this includes another sequence, DQ023431 (the "tag" refers to the uBio record for Apomys datae -- I treat taxonomic names as tags, but that's another story). I could then navigate to DQ023431, and then onto the publication that refers to that sequence, and so on. Very quickly we get a web of data. This is, of course, the vision of the Semantic Web and related projects, such as Linking Open Data. Wonderful idea, it's just such hard work getting there...