Hard for me to know

The article at reference 1, from ‘The Atlantic’, came to my attention a couple of days ago. An article puffing the paper at reference 4; an article which purports to give the coup de grâce to the hypothesis that there is a simple connection between any one of a number of candidate genes and depression. With a leading candidate, showing much promise back in the 1990’s, being something called SLC6A4, involved in the serotonin cycle. Since then, one might say that careers have been built on investigating this connection, with hundreds of papers having been published on the subject.

Given that evidence against has been out there for some time, it is reasonable to ask, on the basis that the present paper has got it right, why it took so long and to close this one down.

One reason is that reading genomes has got hugely quicker and cheaper since the 1990’s and we now have lots of large sets of human gene data connected to the corresponding life data, including data about depression. Data which can be mined for obscure connections between genes and disease (or disorder, to use the term preferred by some mental health professionals).

Another reason is that it has taken a while to realise that few disorders, particular mental disorders, can be put down to a single defect in a single gene. Life is a lot more complicated than that – and so, as it appears to have turned out, the search for a single gene to cause depression was doomed from the outset. Particularly when the choice of the candidate genes might reasonably be described as educated guesswork. Hunches.

Then there are worries that the pressure on researchers to publish flashy results is getting too much. The peer review systems can’t take the strain any more. And worries that fads and fashions rule the research world, in the same way that they rule other worlds.

While I had ambitions to read the story from the horse’s mouth, as it were, that is to say from reference 4, an open access version of which was quickly turned up for me by Bing.

To find that, despite having some background in both this subject matter in particular and statistics in general, while the general drift of the article is comprehensible, I am in no position to comment on its quality or the validity of its conclusions. I quite failed, for example, to work out what Figure 3, snapped above, is about – pretty though it is. Beyond noticing that a gene which might actually have some effect, DRD2, is very close to the bottom. Asking Wikipedia to remind me about p-values was not good enough. Perhaps I should have waded into the supplementary information. 

So I now wonder about the skills which research teams have to assemble to do this sort of work. It does not seem reasonable to expect people who are interested in the subject matter to be statisticians as well, so presumably they have to co-opt one onto the team, or at least have good access to one. Another expense against a research budget which might already be rather stretched.

And funding bodies, made up of relatively lay people, will have to rely on evidence from one expert in making their assessment of a research proposal from another. With the world of research not being that large and with the two researchers involved quite possibly knowing each other. You scratch my back and I’ll scratch yours.

All very difficult.

References

Reference 1: A Waste of 1,000 Research Papers – Ed Yong – 2019. An article in ‘The Atlantic’ online magazine.

Reference 2: https://www.theatlantic.com/world/.

Reference 3: https://en.wikipedia.org/wiki/The_Atlantic. ‘… The Atlantic is an American magazine and multi-platform publisher. Founded in 1857 as The Atlantic Monthly in Boston, Massachusetts, it was a literary and cultural commentary magazine that published leading writers' commentary on the abolition of slavery, education, and other major issues in contemporary political affairs ... After experiencing [difficulties] in the late 20th century, the magazine was purchased by businessman David G. Bradley, who refashioned it as a general editorial magazine primarily aimed at a target audience of serious national readers and thought leaders. In 2010, The Atlantic posted its first profit in a decade ... Its website … provides daily coverage and analysis of breaking news, politics and international affairs, education, technology, health, science, and culture…’.

Reference 4: No Support for Historical Candidate Gene or Candidate Gene-by-Interaction Hypotheses for Major Depression Across Multiple Large Samples - Richard Border, M.A., Emma C. Johnson, Ph.D., Luke M. Evans, Ph.D., Andrew Smolen, Ph.D., Noah Berley, Patrick F. Sullivan, M.D., Matthew C. Keller, Ph.D. – 2019.