statistical significance for small fold changes

[kr067]

Hi everyone,

I have RNA seq data from the same tissue under 2 different experimental conditions and Im trying to find differentially regulated transcripts. I've used Cufflinks/ Cuffdiff to calculate FPKM values and find differentially expressed transcripts. From what I understand, Cufflinks does not run the statistical test in cases where there is low coverage, and many of the transcripts in my samples have low read counts. Also, a couple of my control transcripts have very small fold changes which are not being detected to be significant by Cufflinks. I might be missing detecting small but significant fold changes.

I would really appreciate any ideas on other programs/ statistical tests that I can use to calculate significance in small fold changes.

[steven]

I fear that the problem is more about the data than the software.. powerful statistical methods generally require replicates. If you only have one sample for each experimental condition your differential expression analysis is probably compromised.
You may find additional information if you look for threads with keywords like "replicates", "expression", "statistical", etc. Maybe this one or this one could help.