that the reference database features a satisfactory taxonomic sampling of sequences


that the sequences in the reference database are correctly identified and annotated


that the process of translating the comparison into species names is standardized, universally adopted, and not easily misunderstood
http://plosone.org/article/info:doi/10.1371/journal.pone.0000059#article1.body1.sec1.p1

Dear all,

I have enjoyed reading the article and found it a very good treatment of the problems associated with entries in GenBank, EMBL, DDBJ.

In a traditional/old-fashioned museum (with all the jars with strange looking de-pigmented organisms, pinned butterflies, skeletons etc.) real but dead organisms are reliably stored, categorized and annotated by a curator. If there is any doubt about the taxonomic affiliation one can always examine the specimen. Unlike this traditional museum GenBank, EMBL and DDBJ store only DNA sequences + fragmentary annotation of some organism. Unfortunately in majority of cases this organism, is not linked to a repository (= traditional museum) with the dead organism, or in case of cultures a culture repository (i.e. ATCC). GenBank, EMBL, DDBJ has probably no means how to regulate this problem nor the regulation of the depth of annotation of the sequence as presented by Nilsson et al. in their article because it all stands on the individual researcher that submits the data there.

The point that I would like to highlight is an addition to Nilsson et al.’s “assumptions” in their introduction. One has to also assume that the sequence has been generated correctly and has no errors caused by technical issues (not to be confused by polymorphism).

But really, how reliable are the sequence data in the GenBank, EMBL, DDBJ? Here I would like direct your attention to this excellent article, that is worth reading:

Harris DJ (2003) Can you bank on GenBank? Trends in Ecology & Evolution 18(7) 317-319. http://dx.doi.org/10.1016/S0169-5347(03)00150-2

Anyway, I can only agree that more control over the annotation as well as the sequence is needed from the publishers, repositories and last but not least authors.

Cheers, JanS