Addition of --mergeContext for MethylDackel

[dcarrillox]

Description of feature

Hi,

Have you considered adding the --mergeContext option for the bwameth + MethylDackel workflow? It collapses methylation calls at Cytosine level to CpG site level, and the same goes for the CHG and CHH contexts if --comprehensive is used.

I think it could be a very interesting addition which also would simplify downstream analysis.

Thanks,
Daniel

[sateeshperi]

hi @dcarrillox sounds like a good idea! Do you have any docs reference for the --mergeContext option ?

[dcarrillox]

Hi @sateeshperi, yes! This feature is explained in the Per-CpG/CHG metrics section of the MethylDackel README

[sateeshperi]

the bwameth subworkflow would soon be available as standalone in nf=core/modules repo.
PR here

Would you be interested in contributing to this feature ?

[dcarrillox]

Hi @sateeshperi, yes I can try. Any particular aspect you think I could start with?

[sateeshperi]

hi @dcarrillox could you take a look at the branch that is linked to this issue and if you could verify that the bed graphs now indeed produce merged contexts when running with comprehensive ? that would be gr8

[dcarrillox]

Hi @sateeshperi I am looking at the PR. Very good idea to have the --comprehensive parameter accounting for CHG and CHH methylation contexts, we were planning to implement this as well. However, it should not be bound to the --mergeContext option, they are different things. I may want to have results for CHG and CHH at the cytosine level (as it is provided currently), ie. without the --mergeContext option. We would be happy to implement this as separate parameters, what do you think? Thanks!

[sateeshperi]

@dcarrillox plz go ahead and implement it the right way and submit a new PR. Thanks

[sateeshperi]

or feel free to modify or add on to this PR if that works