Thursday, April 22, 2010

Science, delayed

In our most recent lab meeting,  I presented a recent paper from Science, CKAMP44: A Brain-Specific Protein Attenuating Short-Term Synaptic Plasticity in the Dentate Gyrus.  This was a great, relatively straightforward paper that: 1.) did a proteomics screen to identify a novel AMPA receptor associating protein called CKAMP44; 2.) generated a CKAMP44 antibody 3.) performed immunostaining and northern blots to confirm it was expressed specifically in the brain; 4.) performed westerns to show that CKAMP indeed does associate with AMPA receptors in the brain; 5.) transfected oocytes with CKAMP and measured their modulation of AMPA receptor currents; 6.) generated a KO mouse of CKAMP; 7.) and used the KO mouse to show how CKAMP44 modulates synaptic currents in slices.  For the details, I would recommend reading the paper, since it is relatively straightforward.

Normally, one would think reading such an interesting paper would be a delight.  I, however, was annoyed.  This paper represented years of work by the nine authors.  I suspect that they initially identified the protein 3-4 years ago in the proteomics screen, and confirmed its importance using northern blots/antibodies shortly thereafter.  Yet I had to wait until now to hear about it.  People in the field may have known of CKAMP's existence from conferences, but the information had not disseminated through the community until the paper was published.

Isn't that insane?  That in the age of the internet and instant communication, we as scientists are still waiting months and years to hear about others' research?  Shouldn't we have a better system now?

I have many issues with the current publishing and review system, but the one this paper most applies to is the idea of how journal publishing works. Most of my problems with the system were inspired by Clay Shirky's recent book Here Comes Everybody about how the internet is changing our modes of communication and work.  In one chapter of the book Mr. Shirky described how our models of news is changing.  Before the internet, the model was that journalists would search out interesting stories (as well as be supplied them by publicists or interested parties), filter out the chaff, and publish the newsworthy items; simply put, they filtered then published. This was necessary becase the costs of gathering and transmitting information was high.  For example, if you wanted court information, you had to actually travel to the courthouse, rather than calling them, or looking up the information online.

Now, however, the news model is radically different.  With the internet, everyone has a voice (at least in theory), and can broadcast to their friends what they think is important.  Many news stories now are broken on blogs, and then linked to by other blogs, until they are finally picked up by the major news outlets.  In this model, then, everything is published, and then filtered by users to identify what is important and should be read.

So what does this have to do with science?  The journal publishing system is stuck in the filter-then-publish mode, with editors and reviewers gatekeeping information.  Their jobs (theoretically again) is to verify that scientific findings are true, and of interest.  And to exceed their thresholds for publication, authors need to perform controls and do exciting experiments.

The problem, however, is they don't and can't perform those duties.  It is literally impossible for a reviewer to verify any given work is true, either due to falsification or sloppiness.  Journals are littered with papers that were retracted, or more commonly, never reproduced.  And significance is completely arbitrary, and determined not by journal editors, but after the fact by citations.  I can name many papers in prestigious journals I consider insignificant, and Journal of Neuroscience papers that have been cited one hundreds times (e.g. Rich Mooney's 2000 J Neuroscience paper).

And the cost of this antiquated system is time.  It takes time for scientists to perform all the experiments, beyond the initial, interesting ones; it takes time for authors to put together "stories" (which is an issue for another time), write the paper, and put together pretty figures; it takes time for editors to decide whether to review it, and time for reviewers to pass judgment; and then it takes more time to actually publish it (although this time has lessened with internet publishing).  And if you sum all these time together, you get year long delays between when people do interesting experiments, and the scientific community finds out about them.

Unfortunately, despite my dislike of the current publishing system, I have no simple alternative system.  Whatever the new system entails however, I hope it includes faster publishing times so we can learn of the information faster.