tttianhao / CLEAN

CLEAN: a contrastive learning model for high-quality functional prediction of proteins
MIT License
224 stars 44 forks source link

Some confusion about Fig 2.F in the paper #21

Open sirius777coder opened 1 year ago

sirius777coder commented 1 year ago

Hello,

I have been studying the original paper and came across Fig 2. F, which presents the comparison of CLEAN results with other tools, including BLASTp and ProtInfer. I found it intriguing that the prediction accuracy of BLASTp for EC X.X.X.X is higher than that of EC X.X.X.-, and a similar pattern is observed for ProtInfer. Intuitively, one might expect that if the tool can predict EC X.X.X.X correctly, it should also be able to predict EC X.X.X.- accurately.

Could you please provide some insights or clarification on this observation?