Choice of NEG_TO_POS_RATIO_TRAIN in preparing background region
saninta0212 opened this issue · 0 comments
Hello!
I was wondering if I could get some clarification on how the NEG_TO_POS_RATIO_TRAIN default has been selected. According to the tutorial the default has been set 2, but could I get the reasoning behind why this 2 was chosen and how can I decide if I should choose a different cut-off?
The second question I wanted some clarification on is in this code file here: https://github.com/kundajelab/chrombpnet/blob/master/chrombpnet/helpers/make_gc_matched_negatives/get_gc_matched_negatives.py, the default for "--neg_to_pos_ratio_train" ratio default has been set to 1. Is there a disparity with the tutorial and why was 1 selected here?
If I could also get some explanation on how selecting different values here might affect my downstream analysis that would be great!
Thank you.