Synthesis Begins at Generation: Embedding Route Feasibility, Material Availability, and Process Burden into Molecular Design Download PDF

Generative molecular design can propose chemically valid structures with attractive predicted properties, yet many candidates remain disconnected from a credible path to synthesis. Treating synthesizability as a post hoc filter creates a structural inefficiency: molecular objectives are optimized before route feasibility, precursor availability, operational burden, and planning uncertainty are allowed to influence the design itself. This article develops an original synthesis-constrained design theory in which synthesis begins during molecular generation rather than after it. The proposed construct, termed Route-Embedded Design Viability, represents each candidate as a coupled molecule–route hypothesis characterized by molecular value, step-level reaction feasibility, route topology, context-specific material availability, anticipated process burden, and an explicit uncertainty envelope. The theory distinguishes molecular plausibility from route viability, predicted reaction feasibility from experimental confirmation, catalogue presence from usable material access, and early process indicators from validated manufacturing evidence. It further proposes that molecular value and route viability should be optimized jointly through constrained or Pareto-based search rather than collapsed automatically into a single score. Evaluation should therefore progress from structural integrity and predictive relevance to full retrosynthetic search, route comparison, material verification, process-burden assessment, uncertainty analysis, and expert review of an executable synthesis plan. The contribution is conceptual and methodological rather than empirically validated. Its applicability remains conditional on reaction-data quality, planner configuration, stock definitions, process assumptions, project objectives, and human expertise. By repositioning synthesis evidence as a generative input, the theory provides a bounded basis for designing molecules whose predicted pharmaceutical value is considered together with the practical consequences of making them.