A paper reporting associations between genetic signatures and exceptional longevity, which attracted widespread media attention when published in July last year, has been retracted.
Genetic Signatures of Exceptional Longevity in Humans was published in the well-known journal Science to much skepticism from the statistical genetics community (see Genetic Future article). Having never moved from online access into the full print edition of journal, Science published a formal Editorial Expression of Concern in November 2010, and just this week the authors of the manuscript have fully retracted the manuscript, saying that “technical errors in the Illumina 610 array and an inadequate quality control protocol introduced false-positive single-nucleotide polymorphisms (SNPs) in our findings”. Many of the issues that led to the non-publication of this manuscript and finally its retraction are well articulated by both Daniel MacArthur, in his original blog piece as well as a more recent one, and by the 23andMe research team in their The Spittoon blog.
Comment: Several issues are raised by the ‘publication’ of this article and its subsequent ‘non-publication’. First, it can be argued that the problems with this study should have been picked up in the peer-review process and that the article should not have got as far as online publication. This highlights one of the problems we face with the peer-review process i.e. it’s not fool-proof (although it is probably still the best external reviewing process we have). Hopefully this is something that Science will look into and ask how this manuscript ‘got through’ the process. Second, what part in the process (if any) did publication bias play? These findings were very exciting and would no doubt be of great interest to the research community. However, a closer inspection by the same research community led to widespread skepticism due to errors that were preventable. Third, despite the efforts of the research community to introduce various reporting guidelines, such as STREGA (see previous news), in a bid to highlight the importance of ensuring that research is done to the highest genetic and epidemiological standards, the importance of replication, replication, replication shouldn’t be forgotten (see previous news).
Perhaps most intriguing is how the authors in their retraction “feel the main scientific findings remain supported by the available data” despite “the specific details of the new analysis change substantially from those originally published online to the point of becoming a new report”. Let’s wait and see what gets published.