BMC Genomics has issued an expression of concern for a 2011 paper by a prominent Argentine chemist, Ariel Fernandez, whose work covers several disciplines — “His research spans representation theory in algebra, physical chemistry, molecular biophysics, and more recently, molecular evolution and drug discovery” — and institutions. And therein lies the tale.
Fernandez appeared as the first author of the article, titled “Subfunctionalization reduces the fitness cost of gene duplication in humans by buffering dosage imbalances,” along with a pair of researchers from Taiwan. Fernandez’s affiliations were listed as being with the Instituto Argentino de Matemática “Alberto P. Calderón”, CONICET (National Research Council of Argentina), in Buenos Aires, the Department of Computer Science at the University of Chicago, and the Morgridge Institute for Research, in Madison, Wisc.
According to the abstract:
Recognizing the role of subfunctionalization as a dosage-imbalance buffer in gene duplication events enabled us to reconcile its mechanistic nonadaptive origin with its adaptive role as an enabler of the evolution of genetic redundancy. This constructive role was established in this paper by proving the following assertion: If subfunctionalization is indeed adaptive, its effect on paralog segregation should scale with the dosage sensitivity of the duplicated genes. Thus, subfunctionalization becomes adaptive in response to the selection forces arising from the fitness bottleneck imposed by gene duplication.
The paper has been cited twice, according to Thomson Scientific’s Web of Knowledge.
But as the expression of concern states:
After publication of this article (Fernandez et. al, BMC Genomics 2011, 12:604) it was brought to the Editors’ attention that the data generated by the first author, Ariel Fernandez, seemed anomalous. One of the author’s institutions found that the data were not reproducible from the described methods, but an investigation by the author’s other institution did not find the data or their interpretation suspicious. Given the conflicting conclusions of these investigations, the Editors advise the readers to interpret the data with due caution. We apologize to all affected parties.
So, one of Fernandez’s three institutions, we don’t know which, found cause for concern with his results. Another did not (why only two are referenced here is a mystery). What, we wonder, did Fernandez have to say about all this? If we reach him, or any officials at the various institutions with which he’s affiliated, we’ll update.
Hat tip: Jacek Kominek
On a tangential note: looking at the talk section of the Ariel Fernandez Wikipedia page it has had an investigation all of its own in the past regarding the editing of the page by sockpuppets (hence why the page is now semi protected from edits).
alleged sockpuppets (to clarify)
Will spurious claims of copyright infringement be seen soon? The zeal of certain persons to ensure that their name is not trampled upon is strong.
One of the institutions may be Rice University. Fernandez was professor and Chairman of the Biotechnology Department there for a number of years. The underlying work also seems to be related to dissertations by his advisees:
http://scholarship.rice.edu/handle/1911/13110/browse?value=Fernandez%2C+Ariel&type=author
Rice does not have a “Biotechnology” department, and Ariel Fernandez (who was in bioengineering) was never department chair.
You’re correct. I misread. He was a professor in the department of bioengineering. Sorry.
He also has a retraction for duplicate publication:
http://www.pnas.org/content/103/11/4329.long
There’s also this odd correction for another paper:
http://pubs.acs.org/doi/abs/10.1021/nn302352t
How can anyone publish graphs with no error bars?
There is a big errata in PNAS for duplication http://www.pnas.org/content/108/38/16133.1.full from the same author. Although duplication occurs significantly in both articles at the level of multiple figures and text, one was retracted while the other escaped somehow with just a correction.
Examination of the data is quite revealing (it is provided in a spreadsheet as an additional file).
Here is a plot of the data for humans, which corresponds to panel A of Figure 1 in the paper. Note that there is no fitted line in my plot. What may at first glance look like a fitted line is actually formed by data points that fall right along a line.
Here is a plot of the data for worm, corresponding to panel C of Figure 1 in the paper. Note that there is no fitted curve in my plot. A large mass of data points fall right along a curve (a parabola, in fact).
This can all be seen by careful inspection of Figure 1 in the paper: in panels A, C, and D, most of the density that looks like a fitted line or curve is actually formed by data points.
This pattern seems anomalous indeed.
An expression of concern has been published for another Fernandez paper:
http://www.nature.com/nature/journal/vaop/ncurrent/full/nature13141.html