Forum:Blacklist of questionable genbank entries

Question

[Michael]

Is there a blacklist for incredible/BS genbank entries to use for e.g. blast searches or metagenomics? There could be many questionable or hoax entries in there, or possibly such entries where the taxonomic classification of the sequence is incorrect?

I'll start with this one:

• http://www.ncbi.nlm.nih.gov/sites/entrez?cmd=Search&db=nucleotide&term=AF191073.1&dopt=GenBank

Used to establish the existence of "Stealth virus 1" as a new viral species and "Viteria". Run Blast on the sequence and check the background of the single author W. John Martin, M.D., Ph.D., and his dubious "Center for Complex Infectious Diseases". Edit: IMHO these sequences should be removed from genbank and the patents based on them nullified.

Edit:

See also this: John Martin stripped of his license.

Recommending also this article: http://www.ncbi.nlm.nih.gov/pubmed/15924874 xD

000

From the second publication cited:

With this possible exception, the demonstration of a viral sequence followed by a bacterial sequence clone has yet to be documented. (http://www.ncbi.nlm.nih.gov/pubmed/10331959)

With other words: there is no hard evidence from the sequencing data for a joint occurrence of viral and bacterial sequences occurring jointly together in a single contig; on the other hand the contamination of cell-culture with bacteria or yeast is one of the most common accidents in the wet-lab, in particular if infected material is used in the beginning.

And still, the presence of "Stealth-viruses" and "Viteria" is based on this non-observation and complete sequences of obvious bacterial rDNA origin are annotated as being of viral origin.

[Devon Ryan]

Am I correct in guessing that this person never considered that he just had some bacterial contamination in his cultures?

Edit: Ugg, the person who submitted that has an interview on whale.to. That should result in an autoban from science.

[Michael]

what's whale.to??? edit: ARGH CT alert!

[Devon Ryan]

Yeah, sorry, I should have warned you. Whale.to is synonymous with absurd conspiracy theories and health woo.

[Michael]

exactly my thought, (edit: or that they are just fabricated)

[Istvan Albert]

For a mere $39 we could even read the paper supporting this. You can follow the publication trail in the same journal, Experimental and Molecular Pathology. 1999 -> 2003 -> 2005 -> 2010

[Devon Ryan]

I suspect we'd lose fewer neurons by spending the $39 on booze :p

[Michael]

Fortunately(?) our university has a subscription, out of 30 references, 12 are to, guess, WJ. Martin et al.

[Devon Ryan]

The update that you just made is pretty damning. I'd be amazed he didn't wonder if the findings were wrong, were he not using them to extort money out of sick and desperate people.

[Michael]

Amazingly, he doesn't seem to lie directly in the article, the data is just not backing his conclusion and the abstract is misleading in this respect. Another infamous contribution might be, that these "findings" are used later on to seed FUD about vaccine safety and to support anti-vaccine campaigning.

[Devon Ryan]

It seems he's a typical quack (he does autism woo as well). They don't usually have to explicitly lie, they can just cherry-pick and not use any controls in their experiments. It's unfortunate how nearly impossible it is to get their medical licenses pulled...

[Michael]

This http://www.fda.gov/ICECI/EnforcementActions/WarningLetters/2009/ucm195484.htm might at least have hampered new "research" on stealth viruses.

Original link removed, backup here: https://web.archive.org/web/20170112063325/https://www.fda.gov/ICECI/EnforcementActions/WarningLetters/2009/ucm195484.htm

[Devon Ryan]

Don't hold your breath. The FDA is surprisingly impotent (case in point: Stanislaw Burzynski).

[Michael]

I didn't expect I could better the world that much, as much as I would like to see all quacks give up, but the least thing I can do is help getting rid of these damn spam entries in genbank (end rant).

Answer 1

[Pawel Szczesny]

Not really an answer, just a loose thinking.

Your idea can be tracked back to the very concept of biocuration. Maybe we could restructure the problem in a more efficient way? A blacklist will quickly become obsotele, as both NCBI and Uniprot issue frequent updates that both add and _remove_ data from the databases. Uniprot had described its biocuration procedures here.

Maybe, instead of blacklist, we could issue "community removal requests" or "community correct requests"? A single mail from researcher to NCBI/Uniprot/EnterLargeDatabaseHere isn't likely to have any effect (I've tried, I was kind of dissapointed). But we have a chance (this is a large community after all) to establish a procedure with any of such databases, that would allow them to automatically process our requests.

[Devon Ryan]

This is probably a good thing to try first. In the end, it depends on whether the databases see themselves as performing evidence-based or science-based curation (akin to evidence-based vs. science-based medicine). Databases with evidence-based philosophies will always have bigger problems with BS/woo.

[Michael]

The problem with removal might be that there is an article and patent based on the sequences, if they'd just remove them, the claims couldn't be verified or falsified any more without the supporting sequences.

In general I think your idea is a very good one, (as this is a Forum post I don't expect a complete answer), but this is an excellent starting point for a discussion. I think that this aspect could be a very strong contribution from the BioStar community, I think that the database admins might listen to what we have to say.