There is an overwhelming variety of tools that have been used to predict the binding affinity between a small molecule and a protein. We have gone through the literature and open source code repositories to provide a selection of tools that are ready for you to run on PLEX.
All models we provide are research-grade software and are provided "as-is". No model for this task has yet been demonstrated to generalise well enough to be an alternative to laboratory experiments. We make use of existing, often academic, contributions. Please give credit to the creators of open-source work. We are standing on the shoulder of giants.
At this point in time we are focused on docking. Stay tuned for integrated scoring functions.
Equibind is a very fast, machine learning-based docking tool. The model is less accurate than baseline methods, but orders of magnitude faster.
./plex -app equibind -input-dir testdata/binding/abl
Stärk, H., Ganea, O.-E., Pattanaik, L., Barzilay, R., & Jaakkola, T. (2022). EquiBind: Geometric Deep Learning for Drug Binding Structure Prediction. http://arxiv.org/abs/2202.05146
Base: Gnina (Coming Soon)
Gnina is a sampling and machine learning-based docking tool. Gnina is an implementation of Smina, which itself is a fork of Vina. These tools are considered the current open source baseline.
A McNutt, P Francoeur, R Aggarwal, T Masuda, R Meli, M Ragoza, J Sunseri, DR Koes. J. (2021). GNINA 1.0: Molecular docking with deep learning https://chemrxiv.org/engage/chemrxiv/article-details/60c753ebbb8c1a1a9d3dc142
Diffdock is a machine learning-based docking tool. Diffdock is reportedly faster and more accurate than existing baseline tools.
./plex -app diffdock -input-dir testdata/binding/abl -gpu=true -network=true
Corso, G., Stärk, H., Jing, B., Barzilay, R., & Jaakkola, T. (2022). DiffDock: Diffusion Steps, Twists, and Turns for Molecular Docking. http://arxiv.org/abs/2210.01776