Умный алгоритм устраняет главный тупик при создании сложных противоопухолевых молекул
Molecular conjugates, including PROTACs and peptide-drug conjugates (PDCs), derive their function from the joint behavior of multiple coupled components, yet most generative approaches design these components independently and combine them only after generation. Such staged pipelines ignore cross-component dependencies and often produce conjugates that are chemically invalid, fall outside empirical conjugate distributions, or lose function upon assembly. In this work, we introduce Coupled Objective-Guided Discrete Flows for Molecular Conjugate Assembly (CupOFMoCA), a discrete generative method that enforces cross-component chemical constraints during conjugate generation. CupOFMoCA restricts guided updates to chemically feasible conjugate states and reweights local transitions using molecular design objectives. Across three PDC systems, coupled generation achieved 100% conjugation validity among valid peptides and increased valid yield over matched post-hoc filtering, with comparable predicted binding scores across the full valid sets. In three-seed evaluations of three PROTAC systems, CupOFMoCA achieved the highest assembly validity on two systems and generated more assembly-valid molecules per second than the evaluated baselines. We further evaluated the generated PROTACs with post-hoc DegradeMaster scoring and explicit-solvent molecular dynamics, supporting their predicted activity and conformational properties. Together, these results establish CupOFMoCA as an efficient generator of molecular conjugates, supporting therapeutic development via coordinated control of assembly and molecular properties.
## В чем суть открытия Проектирование бифункциональных гибридных молекул в изоляции часто приводило к тупику: новый вычислительный алгоритм моделирует молекулярный комплекс целиком в едином связывающем кармане, гарантируя точную стыковку с первой попытки. ## Как это устроено: принцип работы Rather than designing functional components independently and stitching them together post hoc, the architecture constructs a coupled discrete flow over the complete assembled state space. During generation, candidate transitions are exponentially tilted toward functional binding objectives while hard structural filters instantly mask out chemically invalid junctions and structural liabilities. ## Результаты и доказательства - **Комплексный скор значимости:** 9.5 / 10.0 - **Уровень доказательности:** code_benchmark - **Бенчмарки и прирост:** Conjugation validity (%), assembly validity (%), predicted binding affinity (pKd), DegradeMaster degradation probability, 3D polar surface area (Ų), ternary clash penalty (kcal/mol), valid molecules per second - **Размер выборки / валидация:** N = 100 generations per target across 3 PDC targets (ERBB2, GNRHR, NRP1) and 3 seeds of 100 candidates across 3 PROTAC systems (EGFR-VHL, BRD9-VHL, CDK6-CRBN); 4 molecular dynamics simulations - **Реализация в открытом доступе:** Да (верифицируемый репозиторий кода) ## Что это меняет на практике Eliminates synthetic dead-ends in therapeutic peptide-drug conjugate and targeted protein degrader pipelines, compressing early-stage oncology drug discovery timelines and lowering chemical синтез costs. ## Границы метода и риски Evaluations rely on surrogate computational affinity oracles and нейросеть activity predictors rather than wet-lab binding assays, full ternary cellular dynamics remain unverified, and CPU-side chemical sanitization creates an inference bottleneck.