How to make DiffBind Run Faster?

Question

[B.N.]

Wondering how to make diffbind run faster. Am on an RStudio server on a linux machine. I set bParallel = TRUE in dba.count(). However, it still seems to be taking a very long time. I notice it's only using 1.5 to 2 GB of RAM max. Is there any way I can further optimize this and make it utilize more RAM? Also, is there any way to see how many cores it's utilizing and make it use more cores too?

Answer 1

[jared.andrews07]

Set the number of cores for your dba object. For example, if your dba object is named tamoxifen:

tamoxifen$config$cores <- 12

Note the parallelization in DiffBind does not work on Windows.

You might also give csaw a try, as it's much quicker and generally more flexible (though as a result requires more adjustments and thought).

[B.N.]

Hello Jared. Thank you for the response. I tried implementing the change and R studio only reports using 1.5 to 2 GB still. Do you have any ideas to make R use more memory for diffbind and would it speed up the process that seems to be taking many many hours?

[jared.andrews07]

Without more code or details, not much more we can provide. Why do you think using more RAM is going to make anything faster?

Regardless, DiffBind is not quick. rtracklayer and csaw can generate read counts much quicker as mentioned.