How to choose peak-associated genes for verification of Chip-seq recovered Motifs?

Question

[yiqiangwang]

Hi all, merry Christmas. I have just done a chip-seq with a protein product of novel gene that possibly acts as a TF (it translocates into nuclear upon some challenges during oncogenesis). At the end, the program returned me ~1500 peaks and possible motifs. About a quarter of these peaks were associated with genes (390 genes). Since this is a novel gene, I totally have no clue about its binding sites, so for all those genes and motifs I have to perform verification experiments. As a start, I plan to choose some most-promising genes to test their expression response to this gene.

Now comes my question: are those genes that associated with more PEAKS are most possible target of this gene?? Or are those genes that contain sequences showing highest homology with MOTIFS are most promising candidate of such target? Any suggestion are greatly appreiciated. Yiqiang

Answer 1

[GouthamAtla]

If the novel TF is regulating some genes, most probably they would be enriched in a pathway relevant to your study. You need to perform a functional enrichment analysis (GO, pathways, gene sets etc) to identify candidate genes to test for.

If the peaks are on the promoters, then its a bit easy as they could be direct targets. If you have many distal peaks, you may want to check if those peaks overlap with any regulatory elements of that tissue and then associate those regulatory elements to most plausible gene.

You have to read papers on how people associate distal peaks to plausible target genes.

[yiqiangwang]

Thanks a lot. Yes I do need to read more references to get a better vision on this topic. Do you happen to have a comprehensive manual or discussion on this to suggest? I am basically an wet-lab scientist so will be slow to find adequate readings on these. Have a wonderful 2018. Yiqiang