Hi, I am working with Single population Rna seq data,each sample has around 500 cells laser captured from antibody stained unfixed tissue. I have too many zeros in the data set and there are some cases where only one sample in a group have zero counts , i think which is affecting my results and analysis. Is doing IQR outliers and performing comparisions advised in this case, because i tried doing outliers and doing comparisions, which gave me pretty good results. How does Deseq2 deal with these type of data. Any suggestions on this is higly appreciated.

Thanks.

If you feel like you need to model the additional zeros not compatible with the NB with high mean and low dispersion for this data, take a look at the scRNA section of the vignette where we provide code for integrating with zinbwave.