NR1H3 p.Arg415Gln Is Not Associated to Multiple Sclerosis Risk

A recent study by Wang et al. (2016a) claims that the low-frequency variant NR1H3 p.Arg415Gln is sufficient to cause multiple sclerosis in certain individuals and determines a patient’s likelihood of primary progressive disease. We sought to replicate this finding in the International MS Genetics Consortium (IMSGC) patient collection, which is 13-fold larger than the collection of Wang et al. (2016a), but we find no evidence that this variant is associated with either MS or disease subtype. Wang et al. (2016a) also report a common variant association in the region, which we show captures the association the IMSGC reported in 2013. Therefore, we conclude that the reported low-frequency association is a false positive, likely generated by insufficient sample size. The claim of NR1H3 mutations describing a Mendelian form of MS—of which no examples exist—can therefore not be substantiated by data. This Matters Arising paper is in response to Wang et al. (2016a), published in Neuron. See also the related Matters Arising paper by Minikel and MacArthur (2016) and the response by Wang et al. (2016b), published in this issue. •We attempt to replicate Wang et al.’s observation that NR1H3 p.Arg415Gln drives multiple sclerosis risk•In a 13-fold larger sample, we find no evidence of association to either MS risk or clinical course•We conclude their result is a false positive due to insufficient statistical rigor and flawed logic The IMSGC find no evidence in >69,000 samples that NR1H3 p.Arg415Gln causes multiple sclerosis in families and determines clinical course, as reported by Wang et al. This refutes the initial claim that NR1H3 mutations describe a Mendelian form of MS.