Abstract 4460: OpencanSARchem: chemistry registration and standardization pipeline for FAIR integration of bioassay data

Abstract Introduction: The rapid growth of medicinal chemistry and bioactivity data has driven AI advancements but inconsistencies in chemical representation across databases make bioactivity-focused data integration challenging. Tautomerism, a key challenge, complicates chemical registration due to the existence of multiple interchangeable molecular forms, significantly impacting database uniqueness and separating bioactivity data across tautomeric forms. We previously developed a standalone, fully deployable solution to address these challenges (Dolciami et al. 2022), now a core component of the canSAR (https://cansar.ai) knowledgebase. However, our solution fell short of achieving FAIR (Findable, Accessible, Interoperable, Reusable) data integration due to its reliance on commercial software. In addition, it was computationally intensive limiting the size of data and rapidity of analysis. We present OpencanSARchem, a fully open-source, RDKit-based, and computationally efficient tool for the drug discovery community. It is 100-fold faster, enabling analysis of larger databases, eliminates reliance on commercial software, and enhances scalability to advance data integration. Methods: OpencanSARchem pipeline involves five steps: (1) assessing input structure for physical consistency, (2) removing irrelevant fragments, solvents, and counter-ions, (3) tautomerically canonicalizing compounds, (4) removing stereochemistry and isotopic labels, and (5) generating abstract representations for hierarchical generalization and grouping of related data across disparate data sources. Tautomeric canonicalization is the key difference between RDKit (OpencanSARchem) and ChemAxon (canSARchem), potentially leading to variations in canonical tautomer selection. We use quantum chemical methods to calculate the Gibbs free energy of tautomer pairs generated by both pipelines, followed by statistical analysis to evaluate the energetic consequences of our utility selection. Results: The new pipeline can process 2.3 million compounds for standardization and generation of canonical tautomers in ∼3 hours. We found that 20% of compounds resulted in different tautomers between ChemAxon and RDKit. The median difference in Gibbs free energy between ChemAxon and RDKit tautomers was found to be approximately 2 kcal/mol. In the context of drug discovery, small free energy differences allow tautomers to coexist in equilibrium, making either structure potentially available for therapeutic activity. Conclusion: To our knowledge, OpencanSARchem is the only fully open-source pipeline addressing all aspects of chemical standardization, including tautomerization, and enabling FAIR data integration, democratizing this method to address the needs of the broader community. Dolciami et al., J. Cheminformatics, 2022, 14, 28. Citation Format: Rezvan Chitsazi, Phillip W. Gingrich, James P. Long, Chong Wu, Bissan Al-Lazikani. OpencanSARchem: chemistry registration and standardization pipeline for FAIR integration of bioassay data [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2025; Part 1 (Regular Abstracts); 2025 Apr 25-30; Chicago, IL. Philadelphia (PA): AACR; Cancer Res 2025;85(8_Suppl_1):Abstract nr 4460.