Dear CNVtools/snpMatrix users, Naim Rashid (to whom we are very grateful) has pointed our attention to a significant bug in the underlying generalized linear model logit C code (shared between CNVtools and snpMatrix). While neither package will crash it is possible that some of the P-values are slightly off. My current data suggest that the effect is unlikely to generate any significant issue but given the potential for errors I recommend all users to upgrade to the latest fixed versions of both packages (now available from bioconductor). Thanks, Vincent [[alternative HTML version deleted]]