Making changes in how to deal with interaction-eQTL.
Instead of processing either standard eQTLs or interaction eQTLs separately, now if the user set interaction=True, the pipeline will run both approaches and save results with different file names.
Thus, we can use the same preprocessed data in multiple runs.
Doing the same thing for trans eQTL (so now you can combine cis, trans, interaction, and trans-interaction in the same run). PS: Not tested for large datasets, trans analysis will probably crash due to GPU RAM.
Instead of processing either standard eQTLs or interaction eQTLs separately, now if the user set interaction=True, the pipeline will run both approaches and save results with different file names. Thus, we can use the same preprocessed data in multiple runs.