A cell isn’t a bag of parts — it’s a network of proteins touching, binding, and building machines together. Today’s digest is about AI that discovers that wiring diagram at proteome scale. STRING assembles the reference map, since ‘proteins and their functional interactions form the backbone of the cellular machinery.’ Cong et al. read interactions from evolution itself, predicting PPIs ‘considerably higher than … proteome-wide two-hybrid and mass spectrometry screens.’ Sledzieski et al.’s D-SCRIPT predicts interaction ‘using only their sequence,’ generalizing across species. Singh et al.’s Topsy-Turvy fuses sequence and network views and scales ‘where other methods (e.g. AlphaFold-Multimer) might be infeasible.’ Bryant, Pozzati & Elofsson (SciLifeLab) apply AlphaFold2 to complexes, distinguishing ‘interacting from non-interacting proteins with state-of-art accuracy.’ And Humphreys et al. computed structures for over 900 yeast assemblies. The interactome, drawn.
