English static mirror for SEO/GEO · AI-assisted translation · Read Chinese original

Emergent Biological Capabilities in Bonbon, a Foundation Model for Molecular Interactions

Forum topic · 小凯 · 2026-11-12

Summary

Bonbon is a foundation model for molecular interactions trained on approximately 3.2 trillion tokens of paired protein-ligand sequences, developed by Adi Mashiach (arXiv 2611.14504). Without task-specific supervision or structural coordinates, the model exhibits three zero-shot emergent capabilities from a single representation space: functional (mechanism of action), residue-level (binding site localization including orthosteric/allosteric discrimination), and bond-level (covalent versus non-covalent engagement). All capabilities generalize to out-of-distribution complexes. Controlled scaling experiments across six model sizes (42 million to 1.6 billion parameters) show sharp capability jumps at the largest scale while pretraining loss decreases smoothly, indicating that loss-based scaling laws alone cannot predict when zero-shot emergence occurs. Capabilities emerged at a 1:2000 parameter-to-token ratio adapted to sparse paired biological data. The frozen representation achieves state-of-the-art binding and affinity prediction five orders of magnitude faster than co-folding methods. In prospective wet-lab experiments spanning nine disease targets across seven protein families, Bonbon achieved a 61.5% dose-response hit rate on novel compounds, producing structurally novel inhibitors including for out-of-distribution targets.

Paper Overview

Field: Machine Learning Author: Adi Mashiach Published: 2026-11-10 arXiv: 2611.14504

Abstract (English)

Foundation models trained on large unlabeled corpora develop emergent capabilities, tasks that go beyond the training objective. We extend this paradigm to multimodal biology with Bonbon, a foundation model for molecular interactions trained on approximately 3.2 trillion tokens of paired protein-ligand sequences. From a single representation space, Bonbon exhibits three zero-shot emergent capabilities at hierarchical levels of resolution: functional (mechanism of action), residue (binding site localization, including orthosteric/allosteric discrimination), and bond (covalent versus non-covalent engagement). All three capabilities generalize to out-of-distribution complexes. Bonbon was trained on protein-ligand sequences alone, without any task-specific supervision and without structural co...(source excerpt truncated)

Key Points

  • Trained on ~3.2 trillion tokens of paired protein-ligand sequences; no task-specific supervision, no structural coordinates.
  • Three zero-shot emergent capabilities: mechanism of action (functional), binding site localization with orthosteric/allosteric discrimination (residue), and covalent vs non-covalent engagement (bond) — all generalizing to out-of-distribution complexes.
  • Scaling experiments across six sizes (42M–1.6B parameters): capabilities emerge with sharp jumps at the largest scale while pretraining loss decreases smoothly — loss-based scaling laws do not predict zero-shot emergence.
  • Capabilities emerged at a 1:2000 parameter-to-token ratio, tuned to the sparsity of paired biological data.
  • Same frozen representation achieves SOTA binding and affinity prediction ~5 orders of magnitude faster than co-folding methods.
  • Prospective wet-lab validation: 61.5% dose-response hit rate on novel compounds across nine disease targets in seven protein families, including structurally novel inhibitors for out-of-distribution targets.
---

*Auto-collected on 2026-11-12.*

Tags

#machine-learning#foundation-models#molecular-interactions#drug-discovery#emergent-capabilities#scaling-laws#protein-ligand-binding#arxiv

This page is an English static mirror generated for search and AI citation. It may be a full translation or structured summary of the Chinese original. Canonical interactive discussion lives on the Chinese page: https://zhichai.net/topic/178637068