Explore the vast range of titles available.

Structural resolution of protein interactions enables mechanistic and functional studies as well as interpretation of disease variants. However, structural data is still missing for most protein interactions because we lack computational and experimental tools at scale. This is particularly true for interactions mediated by short linear motifs occurring in disordered regions of proteins. We find that AlphaFold-Multimer predicts with high sensitivity but limited specificity structures of domain-motif interactions when using small protein fragments as input. Sensitivity decreased substantially when using long protein fragments or full length proteins. We delineated a protein fragmentation strategy particularly suited for the prediction of domain-motif interfaces and applied it to interactions between human proteins associated with neurodevelopmental disorders. This enabled the prediction of highly confident and likely disease-related novel interfaces, which we further experimentally corroborated for FBXO23-STX1B, STX1B-VAMP2, ESRRG-PSMC5, PEX3-PEX19, PEX3-PEX16, and SNRPB-GIGYF1 providing novel molecular insights for diverse biological processes. Our work highlights exciting perspectives, but also reveals clear limitations and the need for future developments to maximize the power of Alphafold-Multimer for interface predictions. Synopsis Based on thorough benchmarking of AlphaFold-Multimer a strategy for structure prediction was developed and applied to 62 protein interactions linked to neurological disease. Six novel protein interfaces were further experimentally corroborated. AlphaFold-Multimer (AF) largely fails to predict structures of interacting proteins involving short linear motifs when using full length sequences. A prediction strategy was developed based on protein fragmentation, which boosts AF sensitivity at costs of specificity. Application of this strategy to 62 protein interactions linked to neurological disease resulted in 18 correct or likely correct structural models. Six novel protein interfaces were further supported by experiments. Based on thorough benchmarking of AlphaFold-Multimer a strategy for structure prediction was developed and applied to 62 protein interactions linked to neurological disease. Six novel protein interfaces were further experimentally corroborated.

Variant effect prediction remains a key challenge to resolve in precision medicine. Sophisticated computational models that exploit sequence conservation and structure are increasingly successful in the characterization of missense variants in folded protein regions. However, 37% of all annotated missense variants reside in 25% of the proteome that is intrinsically disordered, lacking positional sequence conservation and stable structures. To significantly advance variant effect prediction in disordered protein regions, we combined sequence pattern searches with AlphaFold and experiments to structurally annotate 1,300 protein-protein interactions with interfaces mediated by short disordered motifs binding to folded domains in partner proteins. These interfaces were selected based on their overlap with uncertain missense variants enabling reliable prediction of deleterious effects of 1,187 uncertain variants in disordered protein regions. Extensive experimental efforts validate predicted interfaces and deleterious variant effects that were predicted as benign by AlphaMissense. This study demonstrates how critical structural information on protein interaction interfaces is for variant effect prediction especially in disordered protein regions and provides a clear avenue towards its system-wide implementation.

Structural resolution of protein interactions enables mechanistic and functional studies as well as interpretation of disease variants. However, structural data is still missing for most protein interactions because we lack computational and experimental tools at scale. We thoroughly assessed AlphaFold-Multimer accuracy for structure prediction of interactions involving folded domains binding to short linear motifs from the ELM database. The structure predictions were highly sensitive but not very specific when using small protein fragments. Sensitivity decreased substantially when using long protein fragments or full length proteins with intrinsically disordered regions. We delineated a fragmentation strategy to optimize sensitivity and applied it to interactions between proteins associated with neurodevelopmental disorders. This enabled prediction of highly confident and likely disease-related novel interfaces, but also resulted in many high scoring false positive predictions. Experiments supported predicted interfaces between CREBZF-HCFC1, FBXO23-STX1B, STX1B-VAMP2, ESRRG-PSMC5, PEX3-PEX19, PEX3-PEX16, and SNRPB-GIGYF1 providing novel molecular insights for diverse biological processes. Our work highlights exciting perspectives, but also reveals clear limitations and the need for future developments to maximize the power of Alphafold-Multimer for interface predictions.