The next generation of the IUPAC International Chemical Identifier (InChI)

The InChI (IUPAC International Chemical Identifier) is a standardized and machine-readable alphanumeric code that serves as a unique identifier for chemical compounds [1,2] It encodes structural information, allowing for precise identification and communication of chemical structures across databases and research platforms, promoting interoperability in the field of chemistry. That made it become a fundamental tool for the FAIR data handling [3] in chemistry and all related scientific or technical areas that must administer data of chemical compounds. The InChI calculation uses mol files as input format. Based on a multistep process the atom numerations are canonicalized while the structure representation is transferred into a multi layered linear string that shows the connectivity between each atom followed by the number of hydrogens on each atom and other information like the stereochemistry of the molecule or isotopic atoms. The InChI string is hashed to the InChIKey with a fixed length of 27 characters. Because of the fixed size the InChIKey is the ideal key for databases (e.g. in databases with more than 100 million compounds such as PubChem), for the identification of structures within publications, or for Web searches on chemical compounds. InChI applications like the Mixture InChI (MInChI), Nano InChI (NInChI), and Reaction InChI (RInChI) extend the InChI usage into the identification of mixtures and formulations, nano materials and reactions, That make the InChI and its applications to the most important identifier of chemical compounds that is broadly used within the chemical and other related communities. In the last 20 years, the InChI has been primarily developed for the organic chemistry and is now heading towards applications in machine learning, restricted similarity searches and clustering analyses that aids the identification of potential drug candidates and support the understanding of chemical relationships within and between datasets. Due to cheminformatic constraints, metal bonds have been disconnected right from the early days during the InChI processing which renders InChIs for metal complexes and organometallic compounds less meaningful or even meaningless caused by the information loss. Since molecular inorganic compounds gain more and more importance as catalysts, their analysis is of large interest for databases and machine learning application. Hence, we are working on the next InChI generation providing identifiers based on a correct topological description for these substance classes that has to include the complex stereochemistry of inorganics and organometallics as well while InChIs for organic compounds must not change.[4] The base of the new functionality is the InChI Trust move away from the traditional programming to open-source development on GitHub allowing in kind contributions beside the traditional financial trust model under the scientific supervision of the IUPAC.