Transcription factor binding sites in cancer

Question

[caokai001]

Hello everyone!

For Human,We know that there are many snp in cancer cell, we have SNP and ATAC-seq data of this cancer, and obtained TF PWM matrix from jasper.

So how should I guess the binding site of this TF. Meanwhile SNP may also affect co-factor,So TF cann't binding.

I hope to get your advice,Thanks!

[Friederike]

I assume this is what you want to do (correct me if I guessed wrong):

1. Determine regions of accessible chromatin using the ATAC-seq data; These would be represented by the peaks.
2. Determine which TF are likely to bind to those regions. For this, you would have to determine which TF motifs are enriched within the individual peak regions. A popular tool for that AFAIK is HOMER or individual tools of the MEME suite.
3. Once you know which motifs are present in your open regions, you can check whether any of those motifs are affected by the SNPs you have.

Does that help?

[caokai001]

Hello Friederike , thanks for your answer!

I will restate my purpose: I want to get TF binding site from ATAC-seq data in cancer cell ,sush as CTCF binding site in GM12878 cell line, we know that Transcription factors often form complexes that bind to DNA,so I think co-factor is very infortant to TF binding(such as CTCF ). now we only have ATAC-seq ,and SNP data from GM12878.

yes,we will use to MEME suite (AOM) to enrich konwn TF in the vicinity of CTCF。

my idea:

① When we do not consider the influence of co-factor ： we use g2gtools to incorporate SNPs into the reference human genome, https://github.com/churchill-lab/g2gtools ,so we can get GM12878 genome.fa file. then we will use MEME suite(FIMO)to scan Open Region from ATAC-seq to get the CTCF binding site,Although there are some false positives.

② when we consider co-factor: I am confused . because I donot konw how to select region to enrich motif. if I know the co-factor of CTCF ,how to use PWm matrix of CTCF and co-factor , to predict CTCF binding?

That's a complicated question. Thank you for your answer. Friederike