How To Read A Retraction Fuckjillion
October 14, 2009
From the most recent issue of Current Biology:
Fletcher and Rorth have recently discovered that the phenotypes reported in this paper were due to loss of function of both stathmin and the adjacent gene arc-p20, encoding an Arp2/3 component. The mistake was due to use of an incorrect arc-p20 rescue construct as well as mistakes in the subsequent fly crosses. The true stathmin loss-of-function phenotype is quite mild. The authors have confirmed this new result by generating a clean stathmin knockout through homologous recombination. Because the authors cannot cleanly attribute the originally observed effects to Stathmin, they therefore retract the paper. The authors are very sorry for this mistake and apologize for any inconvenience it might have caused.
The retracted paper is here.
Honest mistake?
DrugMonkey 
October 14, 2009 at 4:57 pm
Yep. Some lackey screwed up the crosses. The genetics ended up being more complicated than they should have been, but nobody realized it until someone else couldn’t replicate with correct crosses. Honest mistake, or somebody didn’t train the fly-pusher very well.
LikeLike
October 14, 2009 at 5:11 pm
For a Juicy-er string of retractions, see here:
http://www.nature.com/news/2009/091012/full/news.2009.998.html
Accusations fly, a Science and a JMB paper retracted already, and a Nature and PNAS paper strongly questioned. Questioned as in, the field tried to reproduce your data, and um, no. My take: poor assay, lab really really wanting to see the binding, failure of reviewers to request actual binding data, ignoring flaws (and maybe just full on-fraud)…..
LikeLike
October 14, 2009 at 8:09 pm
You’ll want to head on over to writedit’s place for discussion of that Hellinga disasterMcTrainWreck.
http://writedit.wordpress.com/2009/10/13/hellinga-controversy-expands/
LikeLike
October 14, 2009 at 11:39 pm
Does not surprise me in the least. It’s too easy to work in a fly lab and never check these things, make a lot of assumptions based on markers alone (many of which are wrong to begin with), and nobody requires documentation of any of it in the published papers. We’re just supposed to take their word for it. Just take one look at all the mistakes in FlyBase and that tells you the state of it.
I’m impressed by retractions complete with explanations of what went wrong – somebody had some integrity somewhere (or they were forced to retract it)? Still better than the ones where the former postdoc or student first author is accused of faking data because it can’t be reproduced at all and nobody’s sure where it came from in the first place.
LikeLike
October 15, 2009 at 1:24 am
Honest mistake. Flybase is full of mistakes (it run by volunteers) and many of the flystocks come from not-for-profit running on a shoestring budget stock centers. They provided a vary valid and complete explanation.
LikeLike
October 15, 2009 at 10:01 am
one of my guiding principles is to avoid invoking conspiracy when incompetence explains the result.
LikeLike
October 15, 2009 at 3:53 pm
Many of the fly papers are prone to these mistakes. The flies are usually not fully checked upon arrival, i.e. no Western blot, etc. and the have to a tendency to escape aka fly. Most mutants do not have a single mutation but multiple, making interpretations much more difficult.
LikeLike