Summary of the Webinar Series: What is a Chemical?

Summary of the Webinar Series: What is a Chemical? September 2022-February 2023. Virtual Organizers: Leah McEwen, Ian Bruno, Stuart Chalk and Fatima Mustafa Description Chemical substances touch on all areas of laboratory science and chemistry underlies many critical worldwide issues, including climate, health, food availability and sustainable development. Increased reporting of machine-readable chemical data will support active research in chemistry and related sciences worldwide and will be essential to the development of the interdisciplinary science critical to address the UN Sustainable Development Goals and UNESCO’s priorities around Open Science. IUPAC is the world authority on chemical nomenclature, terminology, and standardized methods of measurement, and is engaging in a concerted effort through collaboration with the broader chemistry and data science communities to translate a range of assets and activities into the digital domain. IUPAC is leading the chemistry WorldFAIR* project-WP3, an EU funded program, aiming to align standards development and implementation with the FAIR data principles. This will facilitate development of guidelines, tools and validation services that support scientists to share and store chemical data in a FAIR manner and support the ability to compile and interpret data across scientific disciplines. In this webinar series, we aimed to: 1. Understand the chemical substance notations used within multiple disciplines (geochemistry, nanochemistry, atmospheric chemistry, environmental chemistry, oceanography, crystallography, etc.). 2. Explore the data resources that are used in these applied areas (life sciences and pharmaceuticals, agriculture and crop protection, dyes and pigments, and machine learning), and understand the current ways of communication and accessing data by other groups. 3. Investigate various digital and machine-readable depictions or notations of chemical substances, reactions, and datasets (InChI, HELM, SMILES, Graphical Representation, Systematic Representation, Media Types, ChEBI, Ontologies4Chem, and notations of Mixtures-Molecules and Complex Substance Schema). Overall, the webinar series highlighted the current status of working with chemical notations, development of digital tools to transform chemical notations into digital entities and ways to implement FAIR data principles across the chemical enterprise. 2. Overview of Webinars 1. "What is a chemical? Handling Chemical Data Across Disciplines" [1] In this webinar, seven experts from diverse backgrounds were invited: Emma Schymanski (Environmental Sciences), Iseult Lynch (Nanomaterials), Lesley Wyborn (Geochemistry), Kerstin Lehnert (Astromaterials), Marie-Lise Dubernet (Astrophysics), Ken Kroenlein (Material Science), and Dylan Walsh (Polymers). The panelists discussed the status of working with machine-readable formats of chemicals within their field and indicated challenges and future needs. The conversation revealed areas of emphasis that users and developers of machine-readable formats should work to advance and to which IUPAC can potentially contribute. Representation and finding chemical data (If we can't represent materials, how are we going to be able to find them?): Identifiers for small molecule chemicals such as InChI and InChIKey are widely used. Similarly for polymers, BigSmiles is in use; however, stochasticity is a challenge to translate into machine-readable format. Polymer properties are based on assemblies, and assembling and processing needs to be known as they affect properties. Moreover, a notation for nanoparticles is still in development. Linking as-synthesized material to various transformed forms (during storage, upon dispersion, in the environment, body etc.) to correlate actual form with effect is needed. This could help geochemists in identifying isotope ratios / crystal phases or domains etc. Real sample complexity (Can we have chemical identifiers to reflect sample complexity?): There are a variety of different types of chemical samples (cosmochemical, ceramic, alloys, nanomaterials, mixtures, etc.) with multiple factors (type, age, source, storage, uniformity, history, etc.) that play a role in sample representation. 2.“What is a chemical? Applying Chemical Data to Industrial Challenges” [2] In this webinar, six experts from multiple applied chemical areas were invited: Lutz Weber (AI & ML), Teodoro Laino (AI & ML), Yannick Djoumbou Feunang (Agir-Chem), Nelson Vinueza Benitez (Dyes and Pigments), Nick Lynch (Pharmaceuticals and life Science), Gunther Schadow (Pharmaceutical-Regulatory). The panelists discussed the challenges and forthcoming regulation requirements for using chemical data in machine-readable formats and provided many recommendations: Need an easy to use more expressive molecular representation, across all nominal identifiers, extending InChI for bigger structures, and making molecules open for everyone! Explore and adopt technologies to aid chemists as they generate data. Define rules for structures, organizing and sharing physical property data with structures. 3. “What is a chemical? User’ Perspectives on Digital Machine Readable Depictions" [3] In this webinar, experts in five various digital machine-readable depictions were hosted: Greg Landrum (InChI), Dana Vanderwall (HELM), Jonathan Goodman (Graphical Representation), Michelle Rogers (Systematic Nomenclature), and Vincent Scalfani (SMILES). The panelists identified some gaps in chemical representation/identifiers such as: Graphical representation is not standardized for small molecules including organometallics, and isomers. The need to define/determine uniqueness, capturing process of macromolecules including glycans, and lipids. Representation of created mixture of unique chemicals and differentiating these to a "natural product" such as canola oil. Collecting critical information to make determinations on consistent representation is challenging. Standardization of description for non-standard structures Variability in interpretation of structures in different jurisdictions The conversation invited users to be aware of common depictions, to use them in different applications and to understand what can and cannot pass from sketching tools into standard formats. 4. “What is a chemical? Innovations in Chemical Descriptions" [4] Experts of five creative descriptions were featured in this webinar: Henry Rzepa (Media Types), Adnan Malik (ChEBI), Oliver Koepler (Ontolgies4Chem), Alex Clark (Mixtures-Molecules), and Ken Kroenlein (Complex Substance Schemas). The panelists identified technical challenges that need to be focused on including: Solubility: need for better solubility models Data capture: FAIR pushing forward, balance with informatics community building connections in parallel Data quality: particularly auto-validating before publication Community consensus: IUPAC chemical classification Many nuanced details arise in real world challenges. ○ What are critical variables in different use cases? ○ How many layers need to be captured in standard data models for interoperability? ○ How many layers arise empirically? How many can be computed? 3. IUPAC Potential Contribution: Overall, the four conversations reflected what the IUPAC can potentially contribute to. Below are some of the suggestions: Representations of challenging mixtures How to capture distributions around samples e.g., polymer length, phases, etc? Education of what machine-readable is and what FAIR is? Modularity of things: the sample space is incredibly complex. IUPAC can identify what the core commonalities are in sample space. Having a high-level controlled vocabulary/machine actionable for the chemical space Unify data models that can be broadly used. There is a need for coordination of activities in enabling the community to not reinvent the wheel e.g vocabulary. IUPAC FAIR Cookbook! reach the base communities at the bench level. Facilitate Interoperability based on sharing across the communities and representations. Supporting interoperability of FAIR, like approach with SMILES+ Expanding InChI to other areas (e.g., organometallics) Education: creating tables and guides of what can be done with different representations, and documenting common use cases Determining what is critical about different classes of entities (small well characterized, what do we need to know about proteins, other macromolecules, nanomaterials) What are the key parameters to capture in representations? May need different approaches for different applications. Classification is another area that intersects with describing chemical moieties. Coordinate common classification schemas for molecular classes and build active chemical and cross-disciplinary ontology community exchange and practice. References [1] https://doi.org/10.5281/zenodo.7259102 [2] https://doi.org/10.5281/zenodo.7259728 [3] https://doi.org/10.5281/zenodo.7435258 [4] https://doi.org/10.5281/zenodo.7683138'Global cooperation on FAIR data policy and practice' (WorldFAIR) has received funding from the European Union's Horizon Europe project call HORIZON-WIDERA-2021-ERA-01-01, grant agreement 101058393