So finally after a lot of discussion about which tools to use for downstream exploratory analysis for Gene Ontology enrichment and ranting on DAVID (for not being updated for long and still being a favourite) it seems that the people behind DAVID really took it by hurt and finally they have come up with the beta release of the update this year few hours back and it is after 6 years, almost over half a decade. I just got the news from@Daniel Swan in another thread and would be curious to see what the community reaction would be. The website says
Release & Version Information
DAVID 6.8 (current beta release) May. 2016
-- The DAVID Knowledgebase completely rebuilt -- Entrez Gene integrated as the central identifier to allow for more timely updates while still incorporating Ensembl and Uniprot as integral data sources -- New GO category (GO Direct) provides GO mappings directly annotated by the source database (no parent terms included) -- New annotation categories -- New list identifier systems added for list uploading and conversion -- A few bugs fixed
Now it will be interesting to see the comparability of the results with other GO tools which have been updated in 2015 and 2016 . Does it also mean that publications putting up DAVID results this year or last year will also re-analyze their data and if there is a huge discrepancy with the latest result will end up coming up with erratums? Or will it be just be treated as a version upgrade and new projects can work and exploit the upgraded version of the tool. I know it is immensely famous among biologists and was for some years not in the list of the Bioinformatics community. What does Bioinformatics researchers have to say about it?
I will start some jobs in coming days do exploit it with some of our published data and compare it with other tools, what does the community think of this upgrade and how are we going to embrace it?
My 2 cents: I have mixed feelings. If they're actually going to support it going forward then great! If there's going to be yet another 6 year gap in updates then they should do everyone a favor and take it offline. I guess time will tell.
The unfortunate reality is that having a .gov extension gave many of us the mistaken impression that the site was actually getting maintained. I shudder to think how many people got less than ideal results because of that.
My favorite plot is this:
In 2015 alone DAVID collected over 5000 citations, yet at that time it contained only 20% of the known annotations. The question: is this a form a scientific neglect or is it misconduct?
Neglect, since it sounds like the group that was originally in charge of it is no longer around.
I feel it was not misconduct , the over-citations of it over the years and being it a favourite for biologist still kept it much popular among most users, even though its usage was going down in dry labs or among Bioinformatics but still wet lab bosses or pure translational labs using bioinformatics expertise always wanted to use for usage in papers owing to its volume of usage. I am happy that we have blogs like Biostars where we discuss about it at length and these days even wet biologists and bosses leading translational labs are taking this into account and try to play with other GO tools. I have to try the new features so I cannot really come up with any comparisons now. But yes the hypothesis generated with old version never seemed too good to be true as I never found such hits with updated versions in other GO tools and my colleagues used to rely on DAVID and took me a while for me to convince them. I have to test the new updated version and compare and then if it is maintained I can then be a bit again among the good books with biologists about their choice of GO tools. But am sure the rebuilt will not hold results that it was giving a few days back and will urge more usage of other GO tools.
The community (us) is partly responsible. There should have been a much bigger awareness of the fact that the resource was not being updated. It was there for all to see on the updates page. This was discussed in multiple threads over last few years also and many of us tried to email DAVID support without any answer.
Whichever government contractor/sys admins set up the website (I recall reading somewhere that that part was outsourced) did a fabulous job though. The site stayed up and available for years without any obvious errors.
I think this is poor practice on the part of the DAVID team. They should have known that their data was getting outdated and display a clear warning that the site was left as a historical relic. If they didn't know then they're probably not trained bioinformaticians and I wouldn't trust their tools. On the users' side, not checking the validity of a resource is also bad scientific practice perhaps caused by poor training.
I get that what you intend to say but I think as others highlighted probably the task was outsourced and to my gut feeling it was being maintained by non-scientists , probably by informatics team with little or no budget past few years which resulted in no such changes or upgrades over half a decade. I guess that over the years of flak at each and every blogs by the community as highlighted by @genomax might have triggered this idea of finally allocating a budget with a new dedicated team followed by a defined upgrade. I am only concerned if this upgrade will also have a proper team who will maintain the site and have someone answer the mails in case of any problems. A dedicated support should also be there with a bit of scientific training. As for the users I cannot say much, even though the resource validity was in question the multitude of papers citing it did not affect much. So whom to blame.
I echo with Devon , the .gov extension infact spoke volumes about the tool and inclined more people to use it despite the fact it was not updated for such a long time , the NIH name weighed it more. But if it does not get the support it is better to take it offline, a website that has been in usage for such a long time over the years and been cited so many times over last 5 years should be well maintained and supported.
The following thread is relevant and I'll link it here for reference. It has quite a bit of back and forth on DAVID:
GO analysis: DAVID vs GREAT vs GOrilla
And of course, as you put it, what about the papers that were written up to last week that used six year old data ... is it reasonable to assume that they would produce the same results? ... gut checks says absolutely not.
And of course ... is it even possible to run old DAVID? So that one can compare it to the new...