FSCseq

FSCseq performs simultaneous clustering and selection of cluster-discriminatory features, using a finite mixture of negative binomial regression model. Inference is done using a classification expectation-maximization algorithm. FSCseq additionally adjusts for effects due to technical variation (like subject-level differences in sequencing depth), and can correct for effects of confounding factors such as batch. Finally, FSCseq can perform prediction of cluster membership on new samples with the learned parameter estimates.