LoGoPPI enables fast and accurate protein protein interaction mapping at scale
Graph-based protein function analysis is powerful, but protein-protein interaction (PPI) networks exist for only a small fraction of animal and plant genomes. We present LoGoPPI, which infers PPIs from sequence by combining bi-encoder global protein representation with local residue-level late interaction. LoGoPPI matches or exceeds state-of-the-art PLM-based cross-encoders while achieving…
Graph-based methods for analyzing protein function are highly effective, yet PPI networks are currently available for only a limited subset of animal and plant genomes. Introducing LoGoPPI, a novel approach that infers PPIs from sequences by merging bi-encoder global protein representations with local residue-level late interaction information.
LoGoPPI matches or surpasses the performance of state-of-the-art PLM-based cross-encoders, while delivering inference speeds up to 1,500 times faster. Its local branch provides residue-level signals that pinpoint interaction interfaces and structurally flexible regions. This remarkable efficiency allows for practical proteome-wide PPI reconstruction at scales previously unattainable for cross-encoder models.
By scaling PPI mapping to tens of thousands of animal and plant species, LoGoPPI offers a scalable framework for conducting comprehensive comparative and functional analysis of protein networks across diverse taxa.
Written by urgent.news from bioRxiv's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.
