Menu

Bad AUC value

2021-10-25
2021-10-25
  • Sherif Elsabbagh

    Hi

    I have a question. I tried to validate the docking experiment by first running docking of actives and decoys against the target protein. I drawed the ROC and the AUC was calculated as 0,65..
    In this case, what can be done to improve this score. How can i manipulate the software to enhance the score

    thanks

     
  • David Koes

    David Koes - 2021-10-25

    That's a great question. As in it is a deep research question and is beyond the scope of this forum. You could apply a variety of consensus scoring techniques. You could apply machine learning. You could use deep learning (e.g. gnina).

     

    Last edit: David Koes 2021-10-25
  • Sherif Elsabbagh

    sorry i didnt know that..
    so you mean to apply other docking softwares and use consensus docking scores..
    but nothing can be changed within smina itself..

     
  • David Koes

    David Koes - 2021-10-25

    You can try increasing sampling by increasing exhaustiveness, but the default value is already pretty good so I would expect diminishing returns - a better use of computational resources would be to evaluate other methods so you can do consensus scoring. You could also try the Vinardo scoring function, it does slightly better at VS.

     
  • Sherif Elsabbagh

    perfect i will try it thanks so much

     
  • Sherif Elsabbagh

    perfect i will try it thanks so much

     

Log in to post a comment.

Want the latest updates on software, tech news, and AI?
Get latest updates about software, tech news, and AI from SourceForge directly in your inbox once a month.