Wednesday, February 18, 2015

How many scientists does it take to screw ... ?

This morning I was flipping through the pages of the week's issue of Nature to see if there was anything interesting. There was. I'll tell you about it later, maybe.

I kept flipping. I found a paper with an intriguing title "New genetic loci link adipose and insulin biology to body fat distribution" [doi:10.1038/nature14132]. The authors found 68 loci in the human genome that are statistically correlated with excess body fat. That's not very interesting because very few of these association studies pan out in the long run. Most of those 68 loci are probably genetic noise.

The end of the paper was somewhat interesting. There seemed to be a lot of authors and affiliations. Let's count the authors 1,2,3,...401! There are 401 authors on this paper and they work at 300 different institutes and universities. The list of authors and affiliations take up three pages in the print edition of the journal!

The next paper has a similar title, "Genetic studies of body mass index yield new insights for obesity biology" [doi: 10.1038/nature14177]. The authors found 97 loci correlated with body mass index (BMI). These loci accounted for ~2.7% of BMI variation.

Not interesting.

There are a lot of authors on the second paper as well. Let's count them: 1,2,3,.....481! There are 481 authors from 347 institutes and universities. The list of authors and affiliations covers almost four pages (!) in the print edition of the journal. Is this a record?

It gets worse. In both lists of authors there are entries under "T" like "The PAGE Consortium" and others. They are marked with double daggers and the footnotes say "A list of authors and affiliations appears in Supplementary Information."

I checked it out. For the second paper there are about 800 additional scientists listed as members of the various consortia.

Am I the only one who thinks this is ridiculous?


27 comments :

  1. No. You are not, but the physicists are worse: I've seen papers with 3000 authors.

    ReplyDelete
    Replies
    1. I didn't realize that there were 3000 physicists with jobs! :-)

      Delete
    2. Yes. Sean Carroll mentioned that in a lecture I recently watched on YouTube. Apparently, any article that uses data from the Large Hadron Collider includes all the scientists involved in its design and construction as co-authors.

      Delete
    3. I know one of those people -- while pursuing his M.Sc., he contributed to the design of some gizmo on the LHC. As a result, his name is on one or two LHC papers -- along with ~150 others. He finds the whole thing a bit of a joke. He now works as a computer security consultant because he can actually make a living that way.

      Delete
    4. Larry: "I didn't realize that there were 3000 physicists with jobs! :-)"

      For that physics paper, under "authors" they should just write "All of them"

      Delete
    5. Bowden: "I've seen physics papers with 3000 authors" L. Moran: "I didn't know there were 3000 physicists with jobs"
      http://sandwalk.blogspot.com/2015/02/how-many-scientists-does-it-take-to.html?showComment=1424275351948#c1257188798150370886


      [Tweeted from DiogenesLamp0]

      Delete
    6. Page charges should have an added charge of $1 per author.

      Delete
    7. Apparently, any article that uses data from the Large Hadron Collider includes all the scientists involved in its design and construction as co-authors.

      This is really ridiculous. Are they also giving co-authorships to the bus drivers who shuttled the lead authors to work, or to the people who did the brick laying when the collider was built? At least that would be consistent!

      Delete
    8. If you want some amusement, search for "Aad G" at Web of Science, and click on "more" when the first paper comes up. I didn't mention his name before posting yesterday as I was in a hurry to go home, and I wanted to check the name before giving it. Anyway, if you're going to be listed with 3000 people in alphabetical order you need to choose your parents carefully, as Dr Aad clearly did. Poor old Dr Zwalinski doesn't have a hope. The list of addresses has 272 entries. Something I didn't notice before is that Dr Aad works at Aix-Marseille Université, so maybe I'll run into him some time.

      Delete
    9. "Page charges should have an added charge of $1 per author." Or even $10 per author. It would hardly hurt real authors at all (especially if the base rate were decreased to compensate) but it would put a stop to a lot of nonsense.

      Mind you, I have a horse in this race, so I'm not neutral. A couple of years ago I counted the number of single-author articles the members of this laboratory had: I came first, with 43, next was my wife, with three, followed by just one other person with one. For two-author papers we came in the same order, with 55, 22 and 11. Some of my colleagues have almost never published a paper with fewer than five authors.

      Delete
    10. Anyway, if you're going to be listed with 3000 people in alphabetical order you need to choose your parents carefully, as Dr Aad clearly did.

      I've told this story before. When I was a graduate student, my supervisor suggested we publish papers with the authors listed in alphabetical order. We did. His last name was "Alberts."

      Delete
    11. @Athel
      But was there a correlation with age and lower numbers of authors, as I suspect there was? It was easier to have single/small number author papers when a single clever experiment was all it took to get in Nature. Those days are past.

      @Larry
      Mathematicans and computer scientists routinely do alphabetical order. It doesn't have the effect you mention because everybody gets that Dr AAA is first arbitrarily. It has the nice effect of eliminating the squabbles over being first or last author that biologists have.

      Delete
    12. But was there a correlation with age...

      Yes, there was. At the same time I did a survey of papers in the Journal of Biological Chemistry: when I started my career the typical paper had two authors, followed by one and three. Today very few have fewer than five.

      Delete
  2. This is how a "consensus" is established; also ridiculous.

    ReplyDelete
    Replies
    1. Again, this is the kind of tripe Barry Arrington writes.

      I'm just sayin'.

      Delete
  3. So every one of these authors gets to cite a Nature paper in their grant renewals, which will be considered a much greater impact in science than the two society journal papers a lab of 4 people publishes the same year.

    ReplyDelete
    Replies
    1. Actually, I was talking about this over lunch with a colleague. I think that putting that Nature paper in your CV when applying for a job could actually be a negative. Same goes for a grant application.

      Delete
    2. As opposed to the world's greatest liars in the name of anti-Darwinism.

      Pest, can you name the most 5 prominent creationists of even 30 years ago? Can you point to ANY of them cited ANYWHERE in the scientific literature in the last 30 years?

      Pest, have you seen the founders of the Intelligent Design movement, lawyer Phillip Johnson and crackpot A. E. Wilder-Smith, cited recently in the scientific literature? Or not just recently, but anywhere? Ever?

      Pest, remember the Wedge Document of 1998? It had 20 year goals, which should be completed 3 years from now. Would you care to copy and paste them in? Or should I?

      Do you even remember what they were? Those Intelligent Design dreams of 17 years ago? They're forgotten, right? Too painful to remember.

      Delete
  4. It would be fun to find out how many of the "authors" have read the paper. I'd be willing to bet that at least 10% don't even know the paper exists.

    ReplyDelete
  5. It's so they can squeeze a couple more songs onto the soundtrack album to make more money.

    Wait, that's why movies have much longer end credits than they used to.

    ReplyDelete
  6. Flagship papers tend to have this type of effect. What I find surprising is that number of authors for a GWAS, which doesn't seem to require a team anywhere close to that number of people, compared to a manual genome annotation for instance (where I'm aware of some projects that can reach these kinds of numbers, because you have specialists for a large number of gene families).

    ReplyDelete
  7. I have been part of a similar consortium, the Human Microbiome Project one, and basically these big papers are required by the funding agencies. If you get funded through a consortium call, you have to publish a paper with all the members before publishing anything with the data on your own.

    ReplyDelete
  8. Not to mention the Foldit paper: http://www.nature.com/nature/journal/v466/n7307/full/nature09304.html

    Among its authors, it lists "Foldit players" - in a supplementary table, these players are identified by their usernames (names like g_s or Nicky666). I'm pretty sure they weren't consulted about the content of the paper.

    ReplyDelete
  9. Paul W Franks (author #394) was robbed! He should've been at least author #393!

    ReplyDelete
  10. According Nature (http://www.nature.com/naturejobs/science/articles/10.1038/nj7539-447a) the "hottest researcher" is one Stacey Gabriel, who has never published a paper as lead author. Dan Graur is not impressed, and neither am I, though she has an h-index of 90.

    ReplyDelete
    Replies
    1. While I agree that her position as "Director of Genomics Platform" seems rather nebulous as to what she is contributing, I really think this idea that middle authors are just padding is rather offensive. I'm a computational biologist, myself, and typically experimentalists come to me to help explain what their data means and whether their results are statistically significant. Because biology values experiment over theory, I typically am not the first (or even second) author on my publications. But I certainly aren't on papers as some sort of sinecure.

      Delete