samsledje / ConPLex

Adapting protein language models and contrastive learning for highly-accurate drug-target interaction prediction.
http://conplex.csail.mit.edu
MIT License
119 stars 32 forks source link

how to use the results.tsv #36

Open Saoge123 opened 3 months ago

Saoge123 commented 3 months ago

Thanks for your impressive work. I run the command of prediction and got the 'results.tsv'. This file contian 3 colunms, the last col should be the distance between ligand and protein embedding.

lig-5092 Protein1 0.0 C43734921 Protein1 0.0 C51120531 Protein1 0.0 C63349888 Protein1 0.004318551 C50397193 Protein1 0.013059744

What is the threshold of this distance? e.g. a value less than that is considered a positive sample, and vice versa.