BioForm-LM: Generative Design of Biologics Formulations via In-Context Learning and Physics-Informed Decoding
Abstract
Abstract Motivation: Formulation design for biologics (proteins, antibodies) is a critical bottleneck in drug development, yet remains largely manual and data-poor. Existing computational tools score or rank a fixed candidate set; no prior work generates formulation recipes de novo . Results We present BioForm-LM, a generative system for biologics formulation design combining mechanistic simulation, in-context few-shot learning, and physics-informed decoding. A mechanistic simulator encoding DLVO colloidal theory and Lumry–Eyring aggregation kinetics generates \(\:{10}^{5}\) synthetic training recipes; a small decoder-only transformer pretrained on this corpus adapts to novel proteins from 3–10 real stability measurements without gradient updates; and a distilled physics critic reranks candidates by best-of-\(\:N\) decoding. On BioFormBench — a curated benchmark of real literature-mined formulations — the model attains perfect calibration on the held-out scFv (Spearman \(\:r=1.0\), \(\:p=0.0\)) from only three in-context examples, and covers 35.3% of formulation space versus 7.5% for random sampling (\(\:4.7\times\:\)). Aggregate leave-one-protein-out recall@10 is 0.167 against 0.080 (random forest) and 0.050 (SVM). Ablations show both in-context adaptation (\(\:-67\text{\%}\) recall) and the physics critic (\(\:-33\text{\%}\)) are necessary. Availability: Code, model weights and the BioFormBench dataset are openly released (see Data and code availability). Contact:…
Read the paper
Related studies
- Preliminary Toxicity Analysis of 3DCRT versus IMRT on the High Dose Arm of the RTOG 0126 Prostate Cancer Trial
- Adaptive Radiotherapy for Head and Neck Cancer: Initial Clinical Outcomes from a Prospective Trial
- American Association of Physicists in Medicine Task Group 263: Standardizing Nomenclatures in Radiation Oncology
- Motion Management Strategies and Technical Issues Associated with Stereotactic Body Radiotherapy of Thoracic and Upper Abdominal Tumors: A Review from NRG Oncology
- High-dose-rate interstitial brachytherapy as a monotherapy for localized prostate cancer: treatment description and preliminary results of a phase I/II clinical trial
- Highly efficient training, refinement, and validation of a knowledge-based plan quality control system for radiotherapy clinical trials
Go deeper in Evidence Hub
Evidence Hub adds a study data sheet (design, sample size, evidence level, key numbers), an AI summary and a chat grounded in the paper's text.