Our E. coli paper was accepted on Nature Communications. We developed an integrative framework that combines structural modeling (PrePPI), protein language models (Topsy-Turvy), and evolutionary information (ZEPPI) to predict PPIs at the proteome scale. Applying this framework to E. coli proteome, we identified hundreds of functionally coherent subnetworks — without using functional annotations, the model reveals meaningful biological predictions, enabling functional annotation for previously uncharacterized proteins. Read more here. 🥂