Deep learning for protein complex structure prediction
Problem
Predicting the 3D structures of protein complexes accurately, a long-standing bottleneck in structural biology.
Why it matters
Protein complex structures underpin understanding of disease mechanisms and drug design.
What the AI did
Deep learning methods now predict protein complex structures with accuracies rivalling experimental methods in many cases, though limitations remain.
Results
Accuracies rivalling experimental methods in many cases, mainly for dimeric complexes; protein disorder and host-pathogen interaction prediction remain unsolved.
Impact
The review documents that deep learning has largely achieved near-experimental accuracy for many protein complexes while outlining remaining challenges.
Record
| Venue | Current Opinion in Structural Biology |
|---|---|
| Published | 2023 |
| Work type | review |
| Authors | Patrick Bryant |
| Citations | 52 (OpenAlex, at import time) |
| References | 44 |
| Model or method | not stated in source |
| Organization | not stated in source |
| Primary source | https://doi.org/10.1016/j.sbi.2023.102529 |
| Open access | hybrid |
| Keywords | protein complexes, structure prediction, deep learning, protein disorder, review |
Importance rating
major — millions of lives, an industry or a field
Protein complex structure prediction is a foundational structural biology task; abstract states accuracies now rival experimental methods in many cases, implying broad scientific impact, though no scale is quantified.