Problem with FPKM values

Question

[msameet]

I recently did a cuffdiff to estimate differential expression under two different conditions. I did not have replicates for the data-sets. I used tophat2 and cuffdiff2 for alignment and differential expression respectively. However, most strangely, when I viewed the alignment on IGV, I can see four genes where the coverage is decent, but the FPKM table shows a value of zero in both conditions. Now I should say, that these genes are tandemly repeated. However, when I tried to look for other repeat regions in the genome, they had some FPKM value assigned to it. I wonder what is going on.

Answer 1

[swbarnes2]

If the reads have a mapping quality of 0, because they are squarely in repeat areas, Cuffdiff might have skipped over it. Maybe other regions are not quite perfectly repetitive.