A variational expectation-maximization framework for balanced multi-scale learning of protein and drug interactions

HIGHLIGHTS

• What: The model showed substantial improvements over the strongest baseline HIGH-PPI, which also incorporates multi-scale information for enhanced predictions, with an increase of 13.81% in the BFS split, 13.06% in the DFS split, 7.69% in the Random split, (Fig 2c and Fig S1). The approach provides an effective perspective for integrating unbalanced multi-scale data. The authors show the efficacy of MUSE across various interaction prediction tasks, highlighting the superiority of integrating atomic structure scale information into molecular interaction predictions. Next, the authors introduce how the authors apply the framework to link . . .
  

   

  

  If you want to have access to all the content you need to log in!

  Thanks :)

  Username or Email

  Password

  Remember me

  Lost your password?

  If you don't have an account, you can create one here.