How can I get a list of transcription factor binding sites from ENCODE or similar for each gene in the human genome in a table similar to this:
gene1, chrid, start, end, tf1, tfbs1-start, tfbs1-end
gene1, chrid, start, end, tf1, tfbs2-start, tfbs2-end
gene1, chrid, start, end, tf2, tfbs1-start, tfbs1-end
gene1, chrid, start, end, tf3, tfbs1-start, tfbs1-end
gene1, chrid, start, end, tf3, tfbs2-start, tfbs2-end
gene1, chrid, start, end, tf3, tfbs3-start, tfbs3-end
gene2...
2 answers
You can use the MEME suite to find transcription factor binding motifs. I used Uniprobe to download my binding motif file. Then I used FIMO (in the MEME suite) to find these motifs across across the genome. There are several outputs to choose from for FIMO and I think you might want GFF3.
It depends on how tech savvy you are...if you have basic linux skills,
You can download the called peaks for each of the TFs in the encode database using wget, and overlap them with the designated promoter/exon intron of each gene using bedTools (in linux) or GenomicRanges (in R).
The other question is whether you would like for the list to be tissue specific or not. If yes, this requires a bit more tweaking, but is easily done in R.
SYDH peaks worked really well for me, but for HAiB, I downloaded the BAM files, merged the replicates, and called my own set of peaks using MACS.
Log in to answer this question.