ROBERT: Bridging the Gap between Machine Learning and Chemistry

David Dalmau, Juan V. Alegre-Requena, Esteban Urriolabeitia

The rapid progress of machine learning (ML) has transformed chemical research. Its integration not only fulfills technological needs but also fosters sustainability through the adoption of digitalized procedures, yielding important benefits for a more environmentally conscious future. Despite this evolution, there are implementation gaps that hinder the widespread adoption of ML protocols among a significant portion of the chemistry community. Herein, we introduce ROBERT,[1] a program designed to make ML more accessible to chemists regardless of their level of programming. This software not only enables researchers to produce results comparable to experts in the field, but also adheres to strict reproducibility and transparency standards.[2] To showcase its capabilities, we conducted benchmarking using six recent ML studies in chemistry successfully demonstrating its effectiveness across databases with varying sizes, ranging from 18 to over 4,000 entries. Additionally, ROBERT introduces innovative end-to-end workflows that enable users to input SMILES and generate ML predictors (Fig. 1).

Figure 1. Example of automated SMILES to predictor workflow.

1. Dalmau, D.; Alegre Requena, J. V., ChemRxiv 2023, DOI: 10.26434/chemrxiv-2023-k994h.
2. Walsh, I.; Fishman, D., Nat. Methods 2021, 18, 1122–1127.

 David Dalmau