#proteinlanguagemodels — Public Fediverse posts

🤖 Are protein language models overhyped without evolutionary context?

🔗 Evolutionary Constraints Improve Protein Large Language Model Predictions for Protein Stability, Binding Regions, and Epistasis. Computational and Structural Biotechnology Journal (CSBJ). DOI: https://doi.org/10.34133/csbj.0226

📚 CSBJ - A Science Partner Journal: https://spj.science.org/journal/csbj

#ProteinAI #ProteinScience #ProteinLanguageModels #Bioinformatics #ComputationalBiology #ArtificialIntelligence #EvolutionaryBiology

🧩 Could structural AI finally decode the complex language of T-cell antigen recognition?

🔗 STAG-LLM: Predicting TCR-pHLA binding with protein language models and computationally generated 3D structures. Computational and Structural Biotechnology Journal, DOI: https://doi.org/10.1016/j.csbj.2025.09.004

📚 CSBJ: https://www.csbj.org/

#StructuralBiology #ProteinLanguageModels #Immunotherapy #ComputationalBiology #Bioinformatics #MachineLearning #Immunology #ProteinEngineering #TCR #HLA #DrugDiscovery #AI

I calculated the embeddings for each possible Alanine/Glycine substitution (X->A and A->G) of DSB-1 and then asked for each position, which other residues are most affected. Again the key phosphoresidue S186 shows up brightly (very highly changed from its original value), and there is suggestion of something going on in the disordered domain (circled). Does anyone working with #proteinEmbeddings or #proteinLanguageModels have any pointers for me?