How to deal with Hi-C replicates from different patients' tissue sample?

Question

[ben.kunfang]

Hi,

I have 8 Hi-C data: 4 of them from primary cancer tissue of four patients and other 4 from drug resistant cancer tissue of other four patients. Now I would like to find the chromatin structure differences between primary cancer and drug resistant cancer. I was wondering if there is need to do some normalization to those 'replicates' to reduce the noise from individual difference and 'batch effect'? If so, is there any recommend algorithm or methods? Thanks in advance!

Best, Kun

Answer 1

[Gordon Smyth]

The Bioconductor package diffHic is specifically designed for this purpose, leveraging the capabilites of the edgeR package. I don't know of other Hi-C methods or tools that can account for batch effects and which properly evaluate variation between biological replicates.

For an example of use see:

Johanson, TM, Lun, ATL, Coughlan, HD, Tan, T, Smyth, GK, Nutt, SL, Allan, RS (2018). Transcription factor-mediated supervision of global genome architecture maintains B cell identity. Nature Immunology 19(11), 1257-1264.

[ben.kunfang]

Thanks for your information. I will try the diffHiC~