Hello! We are having some problems with the differential gene expression analysis of RNAseq data which consists in two different bacterial species in two conditions.

We have done RNAseq and the usual differential gene expression analysis pipelines (Deseq, etc) in one organism under two different conditions and now, we would like to do it between these different bacteria.

We have 2 different bacteria (different species), which have ca. 7000 genes. Around 3500 are orthologous genes. Both bacterial RNAseq (3 replicas for each) were done in the same conditions. We would like to do some statistycal analysis to know if the expression is differential between the orthologous genes of these strains. I have a table with the following information:

GeneID-Species A // Gene Length- Sp A // GeneID-Species B // Gene Length- Sp B // Raw Reads Replica1_Sp A // Raw Reads Replica2_Sp A // Raw Reads Replica3_Sp A // Raw Reads Replica1_Sp B //Raw Reads Replica2_Sp B // Replica3_Sp B

We were trying to use SCBN ("https://bioconductor.org/packages/devel/bioc/vignettes/SCBN/inst/doc/SCBN.html") but the example is done with only one replica for each strain. We think that only one replica per strain is not robust enough. But, as we are not really familiar with the statistics behind the program, we are not sure about it. We have thought about pitting all the replicates against each other and then defining in how many comparisons the gene must be differential in order to be considered trully differential.

Has someone find a solution for this?

I would appreciate any comments or ideas that can help us!