A paper linking a rare gene variant to multiple sclerosis risk has come under scrutiny, Stat News reports.
The paper, which was published in June in Neuron, reported that an exome and Sanger sequencing study of 2,000 patients allowed researchers from the University of British Columbia to find a missense mutation in the NR1H3 gene that they said led to the loss of the LXRA protein and, in turn, to multiple sclerosis.
Right after the paper came out, Daniel MacArthur tweeted that he found it "unconvincing." He further pointed out that the variant the researchers homed in on as disease causing could be found 21 times in the Exome Aggregation Consortium dataset. He and a colleague additionally outlined what they found in a comment at PubMed.
Chris Cotsapas from Yale University and the Broad Institute also noted at PubMed that he and his colleagues attempted to replicate the study using data from the International MS Genetics Consortium, but found that the mutation was only slightly more prevalent in people with MS than without, as Stat News notes. Meanwhile, the University of Pittsburgh's Daniel Weeks and his colleagues commented on the paper itself at Neuron. He tells Stat News that the researchers' statistical calculations were flawed and fixing them leads to a loss of significance.
Senior author Carles Vilariño-Güell from UCB tells Stat News that he is confident in his team's results, noting that he has unpublished data that bolsters the findings. He adds that population-scale genetic data can identify common variants well, while missing rare variants.