search

taehojo / SWAT

SWAT : Sliding Window Association Test
1 stars 0 forks source link

Second section of paper code not included #3

Open Catriona-Miller opened 3 months ago

Catriona-Miller commented 3 months ago

Really like your paper related to this code (Deep learning-based identification of genetic variants: application to Alzheimer’s disease classification) and am interested in trying something similar on my dataset. Just confirming, the code in this repo only covers the first step (splitting the genome into fragments and getting an accuracy for each fragment) not the second part (doing an overlapping SWAT on the selected fragments to get the PIS for each SNP)? Is this code for the second part open source? Would be great to try it out. Thanks!

taehojo commented 3 months ago

Thank you for your interest in our paper. The code in this repo covers two steps. First Step: Splitting the genome into fragments and getting an accuracy for each fragment. Second Step: Selecting top fragments with high accuracy (top_window_indices) and choosing features from these fragments. Training and testing classifiers using the selected features (train_and_test_selected_features). Currently, the part of getting the PIS in the Second Step is not included but we are prioritizing the ranking of SNPs.