A Senior Editor at PLoS ONE has notified me that the paper I first complained about back in September (“A cautionary tale on reading phylogenetic trees“) has been retracted. I still have lots of questions.

As a reminder, the paper, “A tree of life based on ninety-eight expressed genes conserved across diverse eukaryotic species,” inferred a phylogeny of eukaryotes using genes expressed in rice and conserved across 49 eukaryotic genomes, including animals, plants, fungi, and protists. Divergence time estimates among the included species were based on synonymous substitution rates.

The paper first came on my radar because the authors used one of my study organisms, the unicellular green alga Chlamydomonas reinhardtii, as an outgroup. This is weird, because the green algae are known to be closely related to the land plants, of which many were included in the ingroup. It’s a basic principle in phylogenetics that the outgroup cannot be more closely related to any members of the ingroup than the members of the ingroup are to each other. This is true by definition: the outgroup has to be outside of the ingroup.

As I read the paper, more problems became apparent. Some of these were just semantic; frequent readers will know that I’m picky about cladistic terminology. Many, though, were substantive. The outgroup choice simply invalidates everything downstream of that, meaning that every analysis based on the phylogenetic tree is useless. A good tree never grew from a bad seed.

I won’t rehash all of the problems with the paper here. I posted a summary on the PLoS ONE website, and more detailed analyses here and here. A couple of the highlights include a five-fold range of estimates for the same divergence, for example 110 million years to 549 million years for the divergence between plants and animals (none of which is remotely plausible, by the way), and divergence times badly at odds with the fossil record (for example, 124 million years for sea urchin vs. zebrafish, implying that chordates originated in the Cretaceous).

Two of the article’s authors responded to my criticisms with defenses that I found wholly inadequate (“Some responses to ‘A cautionary tale on reading phylogenetic trees‘”). The journal wouldn’t tell me very much, citing confidentiality, but they did confirm that they were “reviewing these concerns.” Which was clearly true, since the paper has now been retracted.

As I’ve said before, the journal failed these authors:

At some stage, the peer review process failed these authors. The problems I’ve identified here and in my previous post could have been fixed, and they should have been fixed as a condition of publication. The best-case scenario here is that the handling editor chose reviewers who were not qualified to evaluate phylogenetic and molecular clock analyses. Even that scenario leaves some blame for the reviewers; any biologist (any educated human, really) should have recognized the absurdity of a mid-Cretaceous divergence between plants and animals.

I would like to know why these problems weren’t fixed before publication. Either they weren’t identified by the reviewers or the handling editor didn’t require the authors to fix them.

The questions I sent to the Publications Manager, and for which I still don’t have answers, were

1. How many external reviewers reviewed the paper?
2. Do any of the external reviewers have expertise in eukaryotic phylogenetics/phylogenomics?
3. Did any of the external reviewers identify what I see as the major problems (outgroup choice, internal contradictions, absurd divergence time estimates, and misinterpretations of the inferred tree)?
4. If so, did the handling editor override the reviewers’ recommendations by not requiring these problems to be fixed?

I give the journal some credit for retracting the paper, and I do appreciate the Senior Editor alerting me of that. We all make mistakes, and what’s important is what we do to fix them. But there’s no reason it should have taken as long as it did, nearly nine months. That’s nine months with nothing from the editors to indicate a problem with the paper, nine months in which 4,000 people read the paper. Thankfully, no one has cited it yet, but who knows how many citing manuscripts are in review, their authors unaware that they’re about to propagate bad information?

The journal has been aware of the problems with the paper for almost all of that time:

.@xprinceps Thank you for raising this to our attention. we are currently looking into the concerns raised.

— PLOS ONE (@PLOSONE) September 28, 2017

It shouldn’t have taken a week to confirm that, yes, Chlamydomonas was an inappropriate choice of outgroup and, yes, the authors really did infer a mid-Cretaceous origin of chordates, and to post an expression of editorial concern. Just “Some concerns have been raised about this paper, and we’re looking into it” would have done.

I’m also not thrilled with the lack of transparency. I understand that the journal can’t (and shouldn’t) give out information that would identify who the reviewers were. None of my questions would have done that, though. “Confidentiality” is a lame excuse for keeping the number of reviewers secret.

I hope the authors will redo their analyses (with a more appropriate outgroup choice) and submit to a new journal, and I hope the new journal does a better job with it than PLoS ONE did.