 This paper presents a new method for accurately modeling the structures of proteins and their complexes using artificial intelligence. The authors used experimental data to develop a candidate-based approach to systematically model novel protein assemblies. They then used a combination of in-cell cross-linking mass spectrometry and co-fractionation mass spectrometry, co-fract MS, to identify protein-protein interactions in the model gram-positive bacterium bacillus subtilis. By controlling for the false positive rate of the predictions, they proposed novel structural models of 153 dimeric and 14 trimeric protein assemblies. Cross-linking MS data independently validated the alpha-fold predictions and scoring. This study also revealed protein-protein interactions inside intact cells, provided structural insight into their interaction interfaces, and was applicable to genetically intractable organisms such as pathogenic bacteria. This article was authored by Francis J. O'Reilly, Andrea Graziardi, Christian Forbrigg, and others. We are article.tv, links in the description below.