Deep Learning Methods for Predicting Structures of Protein Complexes


Most proteins in the body are actually arrangements of multiple protein subunits that interact with each other to form quaternary structures. Examples of these quaternary structures include hemoglobin, DNA polymerase, and ion channels. Until now, no deep learning computational modeling techniques have been capable of effectively predicting the quaternary structure of these protein complexes, opening up a large number of proteins whose structure can be predicted.
Of these quaternary structures, ion channels are particularly significant because ion channels are frequent targets for drug discovery. And due to a large number of ion channels (over 300 types of ion channels just in the cells of the in-ner ear), applying x-ray crystallography and nuclear magnetic resonance to determine all of the structures is not feasi-ble. Therefore, accurate structure prediction using computational modeling will be a substantial advancement for drug discovery.
The first method specifically designed to predicted inter-protein contacts in protein complexes using deep learning and reconstruct quaternary structures from the predicted contacts, this software is currently the best solution for pre-dicting protein quaternary structures in real world environments—namely, when those proteins are interacting with partner proteins.

• Drug discovery
• Discovery of engineered proteins for energy production from algae and microbes

Patent Pending

Quadir, F., Roy, R., Halfmann, R. et al. DNCON2_Inter: predicting interchain contacts for homodimeric and homomultimeric protein com-plexes using multiple sequence alignments of monomers and deep learning. Sci Rep 11, 12295 (2021).
Soltanikazemi, E., Quadir, F., Roy, R.S., Cheng, J. DeepComplex: A web server of predicting protein complex structures by deep learning inter-chain contact prediction and distance-based modelling. Frontiers in Molecurlar Biosciences 8, 827 (2021).

Jianlin Cheng, Professor of Computer Science

Contact Information:

Name : Brett Maland, MBA, JD

Title: Senior Licensing & Business Development Associate, Software & Copyright

Department: MU Technology Advancement Office


Phone: 573-882-1046

Address: 204 Engineering Bldg North Columbia, Missouri 65211