Atlas of predicted protein complex structures across kingdoms - PubMed
5 hours ago
- #AlphaFold
- #Protein Complexes
- #Structural Prediction
- Researchers created an atlas of 1.1 million predicted protein-protein interaction structures using AlphaFold2 and ColabFold.
- The dataset covers proteome-wide interactions across bacteria, archaea, humans, mice, plants, and human-virus pairs.
- The atlas includes 181,671 high-confidence protein complex structures, with 37,855 specifically in the human interactome.
- Structural clustering revealed conserved complex architectures across kingdoms, offering insights into uncharacterized biological functions.
- The study identified candidate viral receptors for Human mastadenovirus A and Papiine alphaherpesvirus 2 via co-immunoprecipitation experiments.
- Comparative analyses with AlphaFold monomeric structures uncovered widespread evolutionary gene fusion and fission events.
- The dataset can enhance protein binding-surface prediction using deep learning, extending utility beyond structural modeling.