identify TFBS for given set of SNPs

Question

[tonja.r]

I have approximately 2000 SNPs and I would like to know if some of the SNPs fall into TFB sites. My idea was to you homer as it is quite simple and needs only bed-file of my SNPs (actually of chip-seq peaks but I guess it is not a problem if I have given it the positions of my snps) as an input and region size. However, for 2000 SNPs and region size 50 I got only 3 known motifs.

Is there any other tool so that I could compare my results?

Answer 1

[Emily]

You could use the Ensembl VEP. Run your list of SNPs against the database then filter for the consequence TF_binding_site_variant.

Answer 2

[Alex Reynolds]

You could use jaspar2meme to convert JASPAR motif data into a form usable with FIMO.

Once you have constructed a list of whole-genome TFBS at the desired FIMO threshold (say, 1e-4 or 1e-5), you could use bedmap to associate those TFBS with your SNPs:

$ bedmap --echo --echo-map-id-uniq --skip-unmapped tfbs.bed snps.bed

which prints a list of TFBSs and their associated SNP IDs, where there are associations.

You could also go the other direction, and get the list of SNP positions and associated TFBSs:

$ bedmap --echo --echo-map --skip-unmapped snps.bed tfbs.bed

Depends on how you want to do the association.