A recent article published in the Journal of Cheminformatics addresses the challenge of confirming the patent status of newly developed compounds, particularly in the pharmaceutical industry, where generating novel structures is crucial for drug production.
The article emphasizes the significance of developing compounds with novel structures for new drug production. It highlights the challenges in confirming the patent status of newly generated compounds, especially with the increasing use of artificial intelligence (AI) in compound generation. The lack of free and user-friendly tools for determining the novelty of compounds in terms of patents is identified as a major hurdle.
Table of Contents
Start with Patents
The study utilizes two public databases, SureChEMBL and Google Patents Public Datasets, to create a reference database of drug-related patented compounds using international patent classification. An exact structure search system is constructed using InChIKey and a relational database system for rapid searching. The research incorporates patent information into the training data for generative AI, enabling the direction of molecule generation by considering the patent status (patented or not).
AI-Driven Drug Development
The article discusses the successful integration of patent status into molecule generation using generative AI. By manipulating the number of generated patented compounds, the method allows for the efficient proposal of novel compounds with high drug-likeness. The study suggests that incorporating patent information into generative AI models can contribute to the development of effective drug compounds.
A new approach
The developed method represents a new approach to molecule generation that considers the patent status of molecules, an essential feature in drug discovery. The incorporation of patent information enables the generation of novel molecules with high drug-likeness, potentially facilitating the efficient development of effective drug compounds.
