Using DNA Methylation Data as Input for ChromHMM

Question

[luminea]

Hello,

I am currently working with DNA methylation and Cut-and-Tag data, and I am looking to add DNA methylation data as an input for ChromHMM . I was wondering if anyone has experience using DNA methylation data as input for ChromHMM.Has any research group attempted this before, and are there any publications on the subject?

Thanks in advance for any insights!

[Ming Tommy Tang]

this tool? https://github.com/malonzm1/LuxHMM or segway https://segway.hoffmanlab.org/doc/3.0.4/index.html

[luminea]

Thank you very much for your response! Regarding LuxHMM, I think it only supports methylation data and not histone modification data from CUT&Tag.

Answer 1

[LChart]

Yes, it's possible to incorporate DNA methylation state into ChromHMM. You can basically use the beta value over a region and threshold it at various levels; ideally 2 (elevated methylation, not) or 3 (hyper, normal, hypo). A separate track encoding the relative abundance of CpGs within the region can also be used.

Few publications really do this, but it's certainly possible to segment the genome in this way. I saw a talk where this approach had been applied to TCGA EPIC profiling; I can't find any publication on the matter, but I believe the tracks combined methylation and mutational profiles.

There are a whole bunch of other methylation HMM implementations as well, you might find something other than ChromHMM that suits your liking.