BiteNet

BiteNet is a computational approach for the large-scale detection of binding sites, that considers protein conformations as the 3D-images, binding sites as the objects on these images to detect, and conformational ensembles of proteins as the 3D-videos to analyze. Particularly, BiteNet is suitable for spatiotemporal detection of hard-to-spot allosteric binding sites.

Currently, there are three models available:  i) protein-small molecule, ii) protein-peptide, and iii) nucleic acid-small molecule binding sites. Select one to use below (otherwise, the protein-small molecule model will be used).

If you use BiteNet please cite:

  • Kozlovskii, Igor, and Petr Popov. “Spatiotemporal identification of druggable binding sites using deep learning.” Communications biology 3.1 (2020): 1-12.
  • Kozlovskii, Igor, and Petr Popov. “Protein–Peptide Binding Site Detection Using 3D Convolutional Neural Networks.” Journal of Chemical Information and Modeling (2021).
  • Kozlovskii, Igor, and Petr Popov. “Structure-based deep learning for binding site detection in nucleic acid macromolecules.” NAR genomics and bioinformatics 3.4 (2021): lqab111.

BiteNet

  • Accepted file types: pdb.
  • Please enter a value between 0 and 1.0.
    Filter out BiteNet's predictions with the probability score lower than the threshold
    Kozlovskii, Igor, and Petr Popov. "Spatiotemporal identification of druggable binding sites using deep learning." Communications biology 3.1 (2020): 1-12.
    Kozlovskii, Igor, and Petr Popov. "Protein–Peptide Binding Site Detection Using 3D Convolutional Neural Networks." Journal of Chemical Information and Modeling (2021).
    The final model will be released upon publication (currently on review)