Weekly AI/ML & Biotech Digest — Nov 9 to Nov 15, 2026
Here is my find of some of the more interesting papers this week. I didn't get a chance to read these papers in depth yet. Hopefully I can find some time soon :-).
- ★ Rapid generation of antibody conformational ensembles using machine learning — We introduce Cadense, a generative framework for antibody ensemble prediction that produces diverse Fv structures from sequence, informed by conformational dynamics.
- ★ Rising cell-free DNA level tracks cancer-associated methylation signal and improves multicancer detection — As total circulating cell-free DNA rises, it becomes enriched in cancer-associated methylation signatures, enabling highly accurate multicancer detection across 14 tumor types. (weekly-pnas, 2026-11-03)
- ★ Structural basis of hypervariable region 1 dynamics on HCV envelope glycoproteins in immune evasion — The C-terminal hypervariable region 1 of the HCV E1E2 glycoprotein dynamically shields conserved neutralizing epitopes, explaining the virus's immune evasion strategy. (weekly-pnas, 2026-11-09)
- ★ Rare recovery of reassortants between contemporary clade 2.3.4.4b H5N1 viruses and human influenza A viruses in vitro — Human-adapted PA or PB2 segments significantly increase polymerase activity in reassorted H5N1 viruses, enabling viable reassortment with human influenza strains. (weekly-pnas, 2026-11-09)
- ★ PharmAgents2: Towards Autonomous End-to-End Drug Discovery — Drug discovery requires continuous scientific decision-making across a vast research space.
- ★ Emergent Biological Capabilities in a Foundation Model for Molecular Interactions — Bonbon was trained on protein-ligand sequences alone, without any task-specific supervision and without structural coordinates (arxiv-digest, 2026-11-10)
- ★ Evaluating the role of pretraining dataset size and diversity on single-cell foundation model performance — The success of transformer-based foundation models on natural language and images has motivated their use in single-cell biology.
- ★ Cross-state representation learning enables prospective discovery of dynamic protein allosteric sites — Allosteric regulation offers multifunctional induction in therapeutic interventions, yet conformational plasticity hinders computational identification of allosteric sites.
- ★ Beyond AlphaFold 3: a standardised benchmark of open-source co-folding models for antibody–antigen complexes, from docking accuracy to interaction specificity — Co-folding models are increasingly used to predict antibody-antigen complexes for epitope mapping, candidate triage and structure-guided de novo design or engineering, yet the latest generation of these models has not been compared under standardized conditions.
- ★ Multimodal cell foundation models require synergistic integration beyond alignment — Richter et al. show that simply aligning different types of biological data does not fully integrate them, revealing when data repeat the same information and when they provide complementary insights. (morning-news-aiml, 2026-11-09)
Comments
Post a Comment