Menu

Extension of Numbers of Conformations and Configurations

Akira Suma
2020-05-27
2020-07-02
  • Akira Suma

    Akira Suma - 2020-05-27

    I would like to ask you concerning how to extend the space
    to be explored by smina; i.e., I want to increase the numbers
    of conformations and configurations of the ligands, that
    are taken as the candidates (of the structures) by the
    algorithms.

    (e.g. I want to increase the number of the positions and
    orientations of the ligands that are arranged in the binding
    pockets of the receptors)

    However, I cannot find the parameters that determine the
    size of the conformational and configurational spaces to be
    explored, in the source code of smina.

    If you would let me know the above-mentioned parameters
    relevant to the size of the space to be explored, and the module
    files where the parameter values are included, I could much
    appreciate it.

    Thank you in advance.

     
  • David Koes

    David Koes - 2020-05-27

    --num_modes sets the number of ligand poses to output (default is 9).

     
  • Akira Suma

    Akira Suma - 2020-06-02

    Hi, David,

    OK. Thank you. But my issue is as follows.

    The structures that I obtained employing smina were not enough
    in terms of the accuracy when I compared them with the experimental
    structures.

    I believe that this problem is derived from the restricted numbers
    of trial structures (i.e., conformations and configurations of ligand
    generated by the default setting of sumina).

    So, I want to perform much more extended conformational and
    configurational searches, by increasing the number of
    generated structures in smina. Which parameters should I modify ?

    Thank you in advance.

     
  • David Koes

    David Koes - 2020-06-02

    You can try increasing the exhaustiveness.

     
  • André Filipe Pereira

    Hi David,
    Congratulations, great software.
    Apart from decreasing the exhaustiveness, is there any way to speed up smina?
    Thank you in advance.

     
  • David Koes

    David Koes - 2020-07-02

    Not really. You can be faster by sampling less which is what exhaustiveness controls. Make sure your --cpu option matches your cores but isn't larger than your exhaustiveness. If you have free cores, run multiple smina jobs.

     
  • André Filipe Pereira

    Is it possible to take advantage of hardware accelerators like GPUs and TPUs offered by Google Colab?

     
  • David Koes

    David Koes - 2020-07-02

    Not with smina. gnina has GPU acceleration but it isn't actually faster.

     

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.