arXiv subjects

Octavian Susanu

Publications and source records attributed to Octavian Susanu.

4 recordsLinked to original sources

arXiv2026

Steerable Single-step Retrosynthesis with Electron-pushing Discrete Flow Matching and Feynman-Kac Steering

Expert chemists perform retrosynthetic analysis in a strategic manner, where each reaction in a route is employed with a specific purpose, combined with mechanistic reasoning represented as electron-pushing arrows to visualize how bonds are formed or broken via electron redistribution. Yet contemporary single-step retrosynthesis models exhibit neither of these properties: their learned reaction distribution is fixed to that of the training data, and reactions are predicted via de-novo generation of the precursors or molecular graph edits, omitting the electron redistribution that governs the chemical transformations. We introduce Verso MAELLE, the first electron-pushing-based single-step retrosynthetic model, representing disconnections as continuous-time Markov chains (CTMC) comprised of edits in the electron space, and is trainable using a discrete flow matching formulation. Being a generative model, Verso MAELLE is optionally steerable at inference time through the Feynman-Kac (FK) framework, using rewards that promote certain types of disconnections. We show that besides in-distribution learning, Verso MAELLE can be steered toward cyclization reactions which are scaffold-constructive but underrepresented in the training corpus, via a ring-breaking reward. As a proof of concept, we show how the electron-pushing reasoning, learned from patented data, combined with a domino incentivizing reward, could steer Verso MAELLE toward creative and strategic disconnections, demonstrated on the key domino sequence that was used to synthesize Punctaporonin U, a complex natural product of the sesquiterpene family.

physics.chem-ph↗

arXiv2026

Mechanistic Reaction Prediction via Discrete Flow Matching on Graph-Structured Electron Occupation

Chemical reactions are fundamentally transformations in electron space, yet most machine learning approaches model them either through de novo generation of product molecules or through heuristic graph edits that operate directly on molecular topology. We introduce MAELLE (MechAnistic Edit fLow-matching on eLectron rEarrangements), which instead models reactions as discrete flow matching over electron occupation vectors. Concretely, we formulate the reactant-to-product mapping as a Continuous-time Markov Chain (CTMC) over the graph-structured integer-valued electron occupation space defined on all bonding, non-bonding, and hydrogen sites. To construct the interpolants between the reactants and products, we generalize the discrete flow matching mixture path to an edit-based formulation, where the electron moves are interpolated using Optimal Transport, yielding a mechanism-like set of moves without elementary step annotations. MAELLE achieves competitive performance on the USPTO-480K benchmark compared with leading reaction prediction models. Beyond in-distribution learning, we evaluate robustness across two out-of-distribution settings - structural complexity and reaction type - and find that MAELLE maintains strong performance where existing methods degrade. Finally, because the learned flow operates over the full electron redistribution, MAELLE naturally recovers mechanistic trajectories that align with known chemistry and can predict side products of a reaction.

cs.AI↗

arXiv2026

Strategy-first synthesis planning for complex natural products

The total synthesis of a complex molecule is among the most demanding intellectual and experimental feats in chemistry: a chemist must plan many steps ahead for how to assemble simple building blocks into an intricate target, devise backup strategies, and anticipate procedural challenges. It is also a profoundly creative activity. For half a century, efforts to automate the retrosynthetic design of natural products and other complex molecules have drawn on catalogued reactions, and the resulting tools now report near-complete success on benchmarks built from that same source. But these tools were shaped to fit benchmarked chemistry, and they falter on many natural products, the frontier of the field, whose densely functionalized, polycyclic architectures demand precisely the inventive chemistry the record contains least. Whether a machine could reasonably design such syntheses like an expert chemist does has remained unclear. Here, we show that SynthEx, an agentic framework built on large language models, plans routes to complex natural products that lie beyond the reach of conventional design algorithms. SynthEx proposes competing strategies, assembles a sequence of routine and key steps into a cohesive route, and critiques and improves its own design; the chemistry it favours is more convergent than existing tools produce, and spans a region of reaction space that catalogue-based tools cannot match. Most notably, in blinded assessments, expert chemists judged its key steps comparable to those of published human syntheses and engaged with them as genuine synthesis plans, a response algorithmic route prediction has not previously accomplished. We release routes to more than a thousand natural products as SynthAtlas, an open, interactive database, and anticipate it will become a shared resource for a collection of complex target molecules that lack existing literature routes.

cs.MA↗

arXiv2026

Agentic generation of verifiable rules for deterministic, self-expanding reaction classification

Computer-assisted synthesis planning breaks target molecules into accessible precursors using large libraries of reaction rules that assign each transformation a deterministic, interpretable label. But chemistry is long-tailed, making manual encoding intractable, and existing tools rely on fixed rulesets that cannot adapt to new chemistries. Here we present a fully automated pipeline in which a multi-agent framework of large language models (LLMs) classifies reactions and writes the rules themselves across 665,901 US patent reactions, generating each rule under a verification loop that tests it against the corpus. It expands a standard taxonomy from 68 to 14,073 classes without human curation. With a lightweight fingerprint classifier, it classifies 97.7\% of unseen reactions, matching a leading proprietary classifier while resolving chemistry more finely and extending on demand to chemistry outside its training distribution. The result is a living reactivity database and a general route to turning generative models into reliable, self-expanding symbolic systems.

cs.AI↗