The second stage of BitSeq embraces the differential expression analysis of transcript expression. Providing expression estimates from replicates of multiple conditions, Log-Normal model of the estimates is used for inferring the condition mean transcript expression and ranking the transcripts based on the likelihood of differential expression.