Re-scan matched sequence with SNP

[Teshome Mulugeta]

Hi,

I found FIMO to be useful and published one article https://bmcgenomics.biomedcentral.com/articles/10.1186/s12864-019-6051-0. Thanks for the tool.

I will continue using it. I have a matched sequence (TGTACAATGGGCAGCAGTCTCT) for a specific motif in position weight matrix in 1200 base pair DNA nucleotide sequence.  I have a single nucleotide variant (SNP) in the match sequence at the 7th position from the left, and the matched sequence is changed to (TGTACATTGGGCAGCAGTCTCT). Now i want to check if that single nucleotide change affects the binding affinity of a transcription factor by re-scanning. My question is:

1) Is it comparable if i rescan (TGTACATTGGGCAGCAGTCTCT) with same position weight matrix? Or should i do it the whole 1200 base pair again? I thought since we identified the matched sequence, scanning the matched sequence with the SNP is OK because when i rescan the original matched sequence (TGTACAATGGGCAGCAGTCTCT) i get identical score and p-value results. 

2) What is the best statistical method to compare the transcription factor binding affinity between the two results? I understand FIMO doesn't understand the biological relevance of the SNP change but I thought if i take the scores difference and the p-value ration will give some useful statistical information. What do you advise?

Best,

Teshome

[cegrant]

FIMO computes each match to a motif PWM independently, so you can simply re-run FIMO on the matched sequences of interest. However, the q-value will no longer be meaningful, since no correction to the p-value for multiple testing is needed.

The raw FIMO score is not calibrated. This means that for a given motif PWM, you can compare wether match scores are smaller or larger, but that's about the extent of it. Presumably higher scores are correlated with higher binding affinity, but it's worth repeating that FIMO doesn't known anything about  biology or chemistry, and it's match scores are only as good as the input motif PWM.

 No comparison be made between scores for different motif PWM, but the p-value of the match is still informative.