I have an RNA-seq dataset with three replicates from plant leaves infected with a pathogen and I would like to analyze the expression of the pathogen genes, if possible. I'm expecting the % of pathogen reads in the data set to be < 10% of the reads. I am aligning the reads to the pathogen and plant genomes separately and running cuffdiff on each genome separately.
I am examining one pathogen gene which, according to the genes.read_group_tracking file has 0 fragments in the first 2 reps and 1 fragment in the third rep. The FPKM for that third rep is 3686.97 and the FPKM in the FPKM in the gene_exp.diff file is 1228. cuffdiff reports a p-value for this gene but I ran it with the default setting for --min-alignment-count which is 10. Why is the FPKM value so high and why is cuffdiff reporting a p-value for this gene? I expect it to be excluded because it is below the --min-alignment-count threshold? Thanks