chromVAR analysis in K562 cells

[RegnerM2015]

Hi @jgranja24 or Sarah Pierce,

Amazing work!

I was interested in your chromVAR analysis in the K562 cell line. I see that for part of your analysis you annotated peaks using the Viestra motif set:

"We subsequently added motif annotations using “addMotifAnnotations”with the
motifs curated and clustered from Vierstra et al."

For the pilot screen, instead of using Viestra motifs to annotate peaks, you used the ENCODE ChIP seq dataset (EncodeTFBS).

Since K562 cells are leukemic and probably have a high mutation rate, do you think the EncodeTFBS annotation would be more accurate in cancer contexts? The motif scanning with Viestra or cisBP motifs via motifmatchr uses the reference genome sequence, which is likely different from the K562 cell line genome. Therefore, if we expect mutations to disrupt motifs, could this be problematic?

Thanks in advance!