{"version":1,"type":"story","url":"https://digestai.news/story/researchers-build-kaml-esm-models-to-predict-protein-electrostatic-pro","json":"https://digestai.news/story/researchers-build-kaml-esm-models-to-predict-protein-electrostatic-pro.json","markdown":"https://digestai.news/story/researchers-build-kaml-esm-models-to-predict-protein-electrostatic-pro.md","slug":"researchers-build-kaml-esm-models-to-predict-protein-electrostatic-pro","headline":"Researchers build KaML-ESM models to predict protein electrostatic properties","summary":"Researchers introduced KaML-ESMs, sequence-based neural networks built on ESM protein language models, to predict protein ionization states. The models were trained on the GAINES-augmented PKAD-3r dataset and achieved root-mean-square errors of about 0.5 units for key residues across four benchmarks.\n\nThe team demonstrated that predicted pKa values can identify functional sites across the human proteome. However, performance degraded on 89 buried engineered mutations lacking evolutionary support. The researchers launched the KaML sequence-based prediction platform for biomolecular simulation and protein engineering.","keyPoints":["KaML-ESMs achieved root-mean-square errors around 0.5 units across four benchmarks","Model performance dropped on 89 buried engineered mutations lacking evolutionary data","The researchers applied KaML-ESM2 to map pKa values across the human proteome"],"whyItMatters":"Using sequence-based AI models to predict protein electrostatics reduces reliance on structure determination, aiding drug discovery and protein design.","category":{"slug":"research","name":"Research","url":"https://digestai.news/category/research"},"entities":{"companies":["Biohub","EvolutionaryScale"],"models":["ESM","ESM2","ESM3","ESMC","KaML-ESM2","KaML-ESMC"],"people":["Marius Wiggert","Shen M.","Dayhoff II G.W.","Kortzak D."]},"firstPublishedAt":"2026-10-24T00:00:00Z","updatedAt":"2026-10-24T00:00:00Z","sourceCount":1,"hasPrimarySource":true,"sources":[{"outlet":"Nature Machine Learning","title":"Predicting protein electrostatics with protein language models","url":"https://nature.com/articles/s41467-026-78406-y","publishedAt":"2026-10-24T00:00:00Z","type":"primary","primary":true,"lead":true}],"sourceNotes":null,"discussions":[],"thread":null,"cite":{"text":"Digest AI, \"Researchers build KaML-ESM models to predict protein electrostatic properties\", 24 October 2026, https://digestai.news/story/researchers-build-kaml-esm-models-to-predict-protein-electrostatic-pro","publisher":"Digest AI","title":"Researchers build KaML-ESM models to predict protein electrostatic properties","datePublished":"2026-10-24T00:00:00Z","url":"https://digestai.news/story/researchers-build-kaml-esm-models-to-predict-protein-electrostatic-pro"},"generatedBy":"Written by Digest AI's editorial model from the linked sources; the sources are the record.","license":"Headlines, digests and key points are written by Digest AI and may be quoted with a link to the story page. Linked articles belong to their publishers. Terms: https://digestai.news/terms#reuse"}