interpreting FIMO results - p value and q value

Question

[Affan]

I used FIMO to scan promoter regions (defined as 10,000 upstream from TSS) only on chromosome 1 of mm9. The results can be viewed here: http://nbcr-222.ucsd.edu/opal-jobs/appFIMO_4.10.01429210665330517302090/fimo.html

I have a couple of questions:

1) For each gene ('sequence name'), it finds a hit on both the positive and negative strand. What does this mean? Can I get rid of one of them and only consider the positive strand? Even the start and end coordinates are the same.

2) Could someone explain to me the p value (and the corresponding q value)? Since the results are sorted by p value and the first result has a p value of 1.64e-6 .. doesn't that mean that that hit is deemed to be a true positive? Why then is the q value so high? Why are my q values so high regardless?

3) I have a control data set. Its a chip seq experiment on the transcription factor of interest. Can I just go through the results from FIMO and determine which results lie within the peaks of the chipseq experiment and as that information to create sensitivity analysis?

Answer 1

[Ram]

1. If the motif you're looking for is its own reverse complement, you'll definitely find hits on both strands between the same co-ordinates.

   EDIT: Just checked your motif (CTATATATAG - this is its own reverse complement, so your results are expected)

2. Not sure about this. Maybe the q-value is low only for really long or really specific/rigid motifs. Maybe your motif matrix is a bit lenient.

3. You could definitely do that. Correlating an analysis of the p-values would also give you a weighted measure of sensitivity.