synthmc: a precedent-based synthesizability score for transition metal complexes

Data-driven methods are increasingly being used for the in silico discovery of transition metal complexes (TMCs), ranging from machine learning models for high-throughput screening and generative models proposing new candidates to evolutionary algorithms for their de novo design. However, computationally discovered or designed TMCs are only rarely realized experimentally, in part because their synthesis is difficult and often not considered during the design process. At the same time, established synthetic accessibility scores for organic small molecules cannot be transferred to TMCs beyond their constituent ligands, as their fragment-based heuristics and molecular representations do not capture the influence of the metal center, coordination mode, and steric effects on coordinability. Here, we present synthmc, a precedent-based synthesizability score for TMCs that combines the precedent of ligands, coordination environments, and degrees of steric crowding in a reference corpus constructed from experimentally reported complexes in the Cambridge Structural Database. Evaluated on random and time-based validation sets, synthmc assigns, on average, lower scores to more recently reported complexes, consistent with their reduced precedent, indicating its ability to identify novel chemistry. We hope that synthmc will help to prioritize synthetically accessible candidates in high-throughput screening and generative design of TMCs and, more broadly, motivate the development of more comprehensive synthesizability assessments for TMCs, for example based on retrosynthetic approaches or expert evaluations.

Authors

Institutions

Publication Details

Journal
ChemRxiv
Published
2026-11-13
DOI
https://doi.org/10.26434/chemrxiv.15012168/v1
Primary Topic
Machine Learning in Materials Science
Type
preprint

Funders

Controls
|||
ALL TIME
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
OCT
NOV
preprint

synthmc: a precedent-based synthesizability score for transition metal complexes

Hannes Kneiding
3 links
ChemRxiv
Machine Learning in Materials Science
preprint

synthmc: a precedent-based synthesizability score for transition metal complexes

Hannes Kneiding
preprint en

Abstract

Data-driven methods are increasingly being used for the in silico discovery of transition metal complexes (TMCs), ranging from machine learning models for high-throughput screening and generative models proposing new candidates to evolutionary algorithms for their de novo design. However, computationally discovered or designed TMCs are only rarely realized experimentally, in part because their synthesis is difficult and often not considered during the design process. At the same time, established synthetic accessibility scores for organic small molecules cannot be transferred to TMCs beyond their constituent ligands, as their fragment-based heuristics and molecular representations do not capture the influence of the metal center, coordination mode, and steric effects on coordinability. Here, we present synthmc, a precedent-based synthesizability score for TMCs that combines the precedent of ligands, coordination environments, and degrees of steric crowding in a reference corpus constructed from experimentally reported complexes in the Cambridge Structural Database. Evaluated on random and time-based validation sets, synthmc assigns, on average, lower scores to more recently reported complexes, consistent with their reduced precedent, indicating its ability to identify novel chemistry. We hope that synthmc will help to prioritize synthetically accessible candidates in high-throughput screening and generative design of TMCs and, more broadly, motivate the development of more comprehensive synthesizability assessments for TMCs, for example based on retrosynthetic approaches or expert evaluations.

ChemRxiv
University of Oxford (GB)
European Commission, Research Executive Agency
Machine Learning in Materials Science
AI Navigator

Ask Laika to Summarize, Analyze, and Connect papers live on the map.

Summarize Papers & Methodologies

Extract key findings, datasets, and comparative methods across publications.

Benchmark Rankings & Visual Analytics

Rank top research institutions, authors, funders, topics, and journals by Field-Weighted Citation Impact (FWCI) and paper volume with instant charts.

Connect Distant Disciplines

Bridge topological clusters on the map to find hidden collaborative intersections.

Laika access coming soon