Patent search for chemical compounds using the “Syntelly” service in the Rospatent digital search platform

This article addresses a pressing issue faced by the Federal Institute of Industrial Property (FIPS) – the need to create a domestic tool for patent search on chemical compounds. The relevance of this task is determined by the complexity and high importance of such searches for patent examination, as well as that many foreign commercial systems leaved Russian market in 2022. Conducting a high-quality patent search in the field of chemistry is a non-trivial task, since the same compound can be represented in documents in various ways: by a structural formula, chemical name, molecular formula, etc., which requires the expert to use specialized tools. The Russian artificial intelligence platform for organic chemistry, “Syntelly”, was chosen as a solution to this problem. As part of a joint project between FIPS and “Syntelly” with the support of the “Skolkovo” foundation, the platform was refined and integrated into the Rospatent Digital Search Platform. A key task was the creation of a tool for extracting chemical information from patent texts and scientific articles. For this purpose, a pipeline based on Named Entity Recognition (NER) technology was developed, enabling the identification of Russian-language chemical names in documents and their conversion into machine-readable structural formulas. To ensure a comprehensive search, modules for substructure search, chemical similarity search, and text search were enhanced. The main result of the work was the successful integration of the “Syntelly” service into the Rospatent Platform, providing the agency’s experts with a modern and independent tool for searching patent and non-patent literature based on chemical structures. The implementation of such a domestic solution helps to strengthen the country’s technological sovereignty. Future plans include further development of the system, including improved structure recognition, implementation of Markush structure search, and reaction search.

1. Lupu, M., Mayer, K., Kando, N. and Trippe, A. J. (2017), Current challenges in patent information retrieval, Springer, Heidelberg, Germany.

2. Ohms, J. (2021), “Current methodologies for chemical compound searching in patents: A case study”, World Patent Information, vol. 66, no. 102055.

3. Fan, V., Qian, Y., Wang, A., Wang, A., Coley, C.W. and Barzilay, R. (2024), “OpenChemIE: An Information Extraction Toolkit for Chemistry Literature”, Journal of Chemical Information and Modeling, vol. 64, no. 14, pp. 5521–5534.

4. Isazawa, T. and Cole, J.M. (2022), “Single Model for Organic and Inorganic Chemical Named Entity Recognition in ChemDataExtractor”, Journal of Chemical Information and Modeling, vol. 62, no. 5, pp. 1207–1213.

5. Wang, X. et al. (2021), “ChemNER: Fine-grained chemistry named entity recognition with ontology-guided distant supervision”, Proceedings of the 2021 Conference on Empirical Methods in Natural Language Processing, November 7–11, 2021, Punta Cana, Dominican Republic.

6. Rajan, K., Brinkhaus, H.O., Agea, M.I. et al. (2023), “DECIMER.ai: an open platform for automated optical chemical structure identification, segmentation and recognition in scientific publications”, Nature Communications, vol. 14, no. 5045.

7. Morin, L. et al. (2023), “MolGrapher: Graph-based Visual Recognition of Chemical Structures”, 2023 IEEE/CVF International Conference on Computer Vision (ICCV), Paris, France, 1–6 October 2023, pp. 19495–19504.

8. Lowe, D.M., Corbett, P.T., Murray-Rust, P. and Glen, R.C (2011), “Chemical Name to Structure: OPSIN, an Open Source Solution”, Journal of Chemical Information and Modeling, vol. 51, no. 3, pp. 739–753.

9. World Intellectual Property Organization (2022), WIPO Standard ST.96, Version 6.0: Recommendation For The Processing Of Intellectual Property Information Using Xml, доступ: https://www.wipo.int/export/sites/www/standards/en/pdf/03–96–01.pdf (Accessed 17 Nov. 2025).