Highlights of ACS Fall 2026, August 23-27, 2026 Part 2 Tue. Aug. 25 2026

This presentation traces the development of patent term from early letters patent through the Statute of Monopolies of 1624, which established a fourteen year term .. by reference to craft apprenticeships and the transmission of technical knowledge. The talk then follows how patent term evolved through statutory reform, industrial change, and eventual international harmonization into the modern twenty year standard. Seen in historical context, this raises questions about whether a single, fixed patent duration meaningfully serves modern chemical innovation [for which “much of the nominal patent term can be consumed before commercial deployment leading to reliance on patent term adjustments, extensions, and regulatory exclusivities.”] or whether greater adaptability is warranted.

Add. info: Lester, S., Zhu, H., 2018. Rethinking the Length of Patent Terms. Am. U. Int’l L. Rev. 34(4), 787-806; Rajec, S.R.W., 2023. Patent Term Tailoring. Ind. L.J. 99(2), 475-514.; DrugPatentWatch, May 11, 2026. 20 Years on Paper: The Real Guide to Drug Patent Duration in the United States. DrugPatentWatch

4521638 - Inherency in chemical patent law: Bridging doctrine and experimental proof. Clinton South, Ballard Spahr LLP

This presentation examines inherency in the context of anticipation and obviousness, focusing on recurring issues in chemical synthesis, formulation, materials processing, and scale-up. Recent litigation involving Exelixis, Inc. and Azurity Pharmaceuticals, Inc. [1] illustrates how impurity levels, degradation pathways, and reaction conditions can become central to the analysis. Similar disputes have arisen in the context of polymers, catalysis, battery materials, and other industrial settings. Guidance from the U.S. Court of Appeals for the Federal Circuit, including Cytiva BioProcess R&D AB v. JSR Corp. [2], reinforces that a missing claim limitation must be necessarily present when the prior art is practiced, not merely foreseeable.

[1] IPR2025-00427 - Azurity Pharmaceuticals Inc v. Exelixis Inc, see for example Petition, Patent Owner’s Preliminary Response, etc.; Johnson, A., Irving, T.(Marbury Law Life Sciences Group), "Essentially Free of” saves Exelixis Claims from PTAB Challenge. LinkedIn. Jun 09 2025, 5 p. (accessed 8.23.26).

[2] Cytiva Bioprocess R&D AB v. JSR Corp., 122 F.4th 876 (Fed. Cir. 2024).; Kass, Lawrence (Steptoe)., Fed. Circ. Inherency Ruling Refines Obviousness Framework. Law 360, January 29, 2025, 5 p.

See also: Goldman, M., Evans, G., Zappia, A., 2015. Inherent Anticipation in the Pharmaceutical and Biotechnology Industries. In:Cold Spring Harb Perspect Med 5, a021006 (13 p.) [115-126]. https://doi.org/10.1101/cshperspect.a021006

4543629 - Patent lexicography: Cautionary tale of “Branched Alkyl” ≠ “Branched Alkyl”. David Rucando, MODALITY LAW PLLC

In recent litigation between Alnylam Pharmaceuticals and Moderna [1], Moderna did not infringe Alnylam’s patents because Alnylam defined its term “Branched Alkyl” as “alkyl in which one carbon atom in the group (1) is bound to at least three other carbon atoms, and (2) is not a ring atom of a cyclic group. Therefore, Alnylam’s claims could not cover Moderna’s cationic lipid SM-102, which only comprised primary or secondary carbons, as used in its COVID-19 vaccine SPIKEVAX®.

[1] Alnylam Pharmaceuticals, Inc. v. Moderna, Inc., 138 F. 4th 1326 – (Fed. Cir. 2025)

Add. info: Rucando, D., Patent Lexicography: When “Branched Alkyl” ≠ “Branched Alkyl” - A Cautionary Tale. Organic Patents®.Blog Apr. 6, 2026; Mueller, Lisa L., 2026. Claim Construction and the Lexicographer’s Trap: Federal Circuit’s Narrow “Branched Alkyl” Construction in Alnylam Pharmaceuticals, Inc. v. Moderna, Inc. Patent Panorama, Mar. 16, 2026

4554349 - Quantum computing and IP: An overview of the evolving patent landscape. Yoko Bian, Goodwin Procter LLP

This presentation gives a broad overview of the patent strategies of leading companies and research institutions across the principal quantum computing modalities, including superconducting qubits, trapped ions, photonic systems, and topological approaches, with particular attention to the role of novel materials, including nanomaterial-based components and cryogenic fabrication processes, in shaping claim scope and freedom-to-operate considerations. On the software side, we survey IP activity in quantum error correction, variational quantum eigensolvers for computational chemistry, and optimization algorithms, highlighting the unique patentability challenges posed by algorithm-hardware co-design

Add. info: Aboy, M., Crespo, C., Minssen, T., 2026. Mapping the Patent Landscape of Quantum Technologies: Evolving Patenting Trends and Policy Implications (2025 Update), in: Aboy, M., Corrales Compagnucci, M., Minssen, T. (Eds.), Quantum Technology Governance I: Law and Regulation. Springer Nature, Singapore, pp. 125–147. https://doi.org/10.1007/978-981-95-8371-3_4; EPO, OECD, December 2025. Mapping the global quantum ecosystem. European Patent Office. 168 p. (See Ch. 2. Mapping quantum technologies, pp. 25-26, Ch. 3. Quantum patent landscaping and Annex A.1 Quantum patent cartography: supplementary material, pp.136-139); QuIC, February 2026. A Portrait of the Global Patent Landscape in Quantum Technologies White Paper, The European Quantum Industry Consortium (QuIC), 37 p. (See Appendix A: Search Methodology, pp.31-36) ; EPO, January 2023. Quantum computing Insight report, Technology insight reports. EPO. 42 p. (See Ch. 2 Methodology and sources of patent information used, pp.10-13 and Ch.3 Analysis, p.14-39)

4523128 - Understanding FDA's New Drug Exclusivity Norman Schmuff

This session examines the regulatory distinction between New Molecular Entity (NME) [an administrative classification] and New Chemical Entity (NCE) status [defined in regulation] and explains how these designations are applied to eligibility for five year exclusivity under section 505 of the Federal Food, Drug, and Cosmetic Act. The presentation explores how FDA evaluates active moieties across challenging chemical scenarios, including esters, non ester prodrugs, stereoisomers, metal complexes, and coordination compounds. Selected regulatory and judicial precedents illustrate how small chemical differences can lead to materially different exclusivity outcomes

Add. info: FDA CDER, Feb. 2022. NDA Classification Codes.MAPP 5018.2. FDA, 10 p.. (NME, p.1, NCE vs, NME, p. 2, footnote 2. Esters and metal-containing substances, p.8); CDER Development of New Stereoisomeric Drugs. Guidance Document, May 1992; O’Brien, G., Feb. 23, 2023. Regulatory Exclusivity for Novel Drugs and Biologics.FDA Lifecycle Management Webinar. Mayer Brown, pp. 5-29; Oronsky, B., et al.., 2023. Patent and Marketing Exclusivities 101 for Drug Developers. Recent Patents on Biotechnology 17(3), 257–270. https://doi.org/10.2174/1872208317666230111105223; Nemlekar, A., Stakleff, N.N., Miller, R.A. (Troutman Pepper Locke), (Sep. 7, 2016). FDA Is Evolving on Qualifications for ’New Chemical Entity. Law 360.

CINF Skolnik Award Symposium: Antony Williams, 8/25, AM

4523043 - Cheminformatics for all shapes and sizes (or: It is a mistake to think you can solve any major problems just with potatoes), Alex Clark, Collaborative Drug Discovery

This presentation will describe contemporary methods for representing entities such as inorganics, polymers, synthetic peptides, functionalized RNA and antibody-drug conjugates. By layering on top of the fundamental connection table it is possible to satisfy many different biology and materials use cases without losing access to chemistry algorithms that work on atoms and bonds. Some of the recent developments for chemically-aware biologics will be showcased.

Add, Info: Registration Beyond Sequences: Why You Need Chemically Aware Biologics. CDD Blog, Aug. 17, 2026; Clark, A.M.Synthetic Peptides, Nucleotides, Antibodies and Their Drug Conjugates: For Biologists, Chemists and Computers. ibid., Nov. 14, 2025; Chemically-Aware Antibody and Antibody Drug Conjugate (ADC) Registration [Vault Snack #31, Sep. 29, 2025]; Updates: Automated Antibody Markup (6/2026); siRNA Composition and Registration (3/2026); Macromolecule Registration w/ custom monomers (1/2026); Chemically-Aware Registration for Antibodies and ADCs(Dec 11, 2025 webinar, @2:27-15:21); User-defined custom monomers (10/2025); Chemically Aware Antibody Registration (9/2025); Enhanced biologics and bioconjugates registration (Sep 4, 2025 webinar, @1:10-20:51); Modified Bioconjugates (Structure Editor, Registration) (2/2025); Biologics (Bioconjugate, siRNA and Antibody registration datasheets); CCD Knowledgebase: Macromolecules (Macromolecule Drawing Tool; Custom Monomers or Building Blocks; siRNA,ADC, Antibody, Peptide or Protein registrations); Peptide Registration (w/ chem.. modifications); Chen, W.L., Leland, B.A., Durant, J.L., et al., 2011. Self-Contained Sequence Representation: Bridging the Gap between Bioinformatics and Cheminformatics. J. Chem. Inf. Model. 51(9), 2186–2208. https://doi.org/10.1021/ci2001988

4524702 - Open chemical data in the age of AI. Noel O'Boyle, European Bioinformatics Institute

This presentation explores the role of EMBL-EBI’s chemical biology resources, ChEMBL, SureChEMBL, ChEBI, UniChem and OPSIN, in enabling data-driven research and next-generation AI applications across chemistry and drug discovery. The talk will discuss how open chemical databases and tools are evolving to meet the needs of AI-driven science, the challenges of structure representation, data quality and provenance, and the importance of community standards and open collaboration.

Add. info: O’Boyle, N., (11 June 2026) Introduction Day Two. 2nd ChEMBL User Group Meeting, EMBL-EBI, Hinxton, UK, 10 -11 June 2026. 38 p. (See “Curation of high-quality ML-ready data requires resource”, p. 6-7; OPSIN, UniChem, ChEBI, pp. 9-24); Bosc, N., O’Boyle, N., 2026. (10 June 2026) SureChEMBL Your next source of research data., ibid. 18 p. video, 21:30; Hanser, Thierry. (10 June 2026). Using ChEMBL in an agentic AI framework [PaperClip], ibid. 45 p., video, 20:08, an example of the PaperClip Chat by GXL session: “Find all approved kinase inhibitors that target EGFR with IC50 less than 100 nM”(Aug. 25, 2026)

CINF Skolnik Award Symposium: Antony Williams, 8/25, PM

4538125 - Improving the chemical entity integration in PubChem: introducing RefChem, Evan Bolton, Qingliang Li, Sunghwan Kim, Paul Thiessen, Jian Zhang, Tiejun Cheng, Asta Gindulyte, Siqian He, and Bo Yu, National Center for Biotechnology Information

How a chemical substance is defined across PubChem multiple scientific domains is varied, even among curated and authoritative data sources. To address this challenge, we introduce "RefChem", the PubChem Reference Collection, designed to disambiguate and semantically organize chemical substances using structures, names, and identifiers for high quality chemical information linking. As a part of this, the "PubChem Reference Collection" and how it is used to construct "RefChem" will be discussed.

Add. info: RefChem (data organization); RefChem Dataset [932,904 records as of 8/24/2026, Search Fields: Preferred Chemical Name; Synonyms; Other CIDs; Preferred InChI Key; Other InChI Keys; Filter by Identifiers: RefChem (REFSID) and Substances (SID)]
Note: Compounds/Substances listed in RefChem generates additional synonyms to Pubchem Compound and Pubchem Substance records: “RefChem REFSID”, for example, ‘RefChem:1092876’, and could be search in both databases by keyword RefChem in the Synonym field.

Examples of compounds from RefChem which is linked to patents in Pubchem are given in the Footnote 1 at the end of Part 2(contd).

CINF Chemical Information Across the Chemistry Enterprise 8/25 2:00-4:00 PM CDT

4542390 - InChI today: is there anything it cannot do? | Poster Board #1389. Jonathan Goodman, University of Cambridge, and Gerd Blanke, InChI Trust

The same presentation as the Poster Board #1391 presented at the CINF Sci-Mix, Mon. 8/24 8:00 PM - 10:00 PM CDT

Add. Goodman, Jonathan Subcommittee / Working Group Update in Progress and Plans for InChI, March 2024, 40:04- 1:14:19

CINF Agentic Generative AI and Molecular Data Representation for AI, 8/25 PM

4537916 - ANNalog: an Artificial Neural Network (ANN) for generation of medchem analogs, Noel O'Boyle, EMBL, Wei Dai and Arianna Fornili, Queen Mary University of London, Jonathan Tyzack, Nxera Pharma UK, Chris de Graaf, Structure Therapeutics USA Inc

Here we describe ANNalog, an artificial neural network (ANN) that given a SMILES string can generate (1) structurally similar analogues involving minor modifications, and (2) structurally distinct yet chemically relevant analogues (scaffold hopping). A Seq2Seq transformer model was trained on the [608,642] pairs of molecules from the same ChEMBL bioactivity assay. When the generation process was constrained using ANNalog’s prefix control feature [excluding original scaffold – ab], approximately 25% of the known scaffolds from the patent set were successfully rediscovered by the model, illustrating enhanced performance under user-guided conditions.

Add. Info: Dai, Wei (10 Jun 2026). ANNalog – a generative model for medchem analogues. 2nd ChEMBL User Group Meeting, EMBL-EBI, Hinxton, UK, 10 -11 June 2026 (Slides, 18 p., video, 18:15);

Dai, W., Tyzack, J.D., Fornili, A., de Graaf, C., O’Boyle, N.M.. (Apr. 15, 2026) ANNalog — Generation of MedChem-similar Molecules. J Cheminform 18(1), 53. https://doi.org/10.1186/s13321-026-01186-6;

ANNalog testing colab. Google Colab, https://colab.research.google.com/drive/1aJhaBOG7xuYFwMGzfUmbMsLe8T462Ptc#scrollTo=Ss1QOzXjzKSP; DVNecromancer, ANNalog package, https://github.com/DVNecromancer/ANNalog; O’Bloyle, N. ANNalog, a generative model for MedChem analogs. Noel O’Blog.Apr. 16, 2026; O’Boyle, N., The wrong type of SMILES. Noel O’Blog, Jun. 29, 2026; iwatobipen [Serizawa, Takayuki] Genearate molecule with Annalog. Is life worth living? Blog. Jun. 27, 2026

4542310 - InChI and Stereochemistry: what can it do now, and what will it do next? Jonathan Goodman, University of Cambridge, Gerd Blanke, InChI Trust, Sonja Herres-Pawlis, RWTH Aachen, Andrey Yerin, ACD/Labs [Confience]

In this presentation, we will outline the stereochemical features that InChI handles effectively and identify the categories that remain challenging or are currently unsupported. The ongoing developments aim to incorporate atropisomerism, enhanced stereochemical descriptors, and the complex stereochemistry commonly encountered in inorganic and organometallic systems.

Add. Info: Extended Stereochemistry InChI Project Group (Chair Andrey Yerin, enhanced stereo marks; tropisomers; longer allenes; etc.); Herres-Pawlis, S., (Apr 07 2026). Technical InChI Meeting, March 25–26, 2026. LinkedIn.(Priorities: inorganic stereochemistry, advanced organic stereochemistry (MDL enhanced stereochemistry, atropisomers)); Blanke, G., Herres-Pawlis, S., (14 August 2025). Shaping the Future of InChI: Technical Meeting at RWTH Aachen, 22-23 July 2025. InChI Trust. (Stereochemistry in Molecular Inorganics; Atropisomers: “trust the input”); InChI Workshop on Inorganic Stereochemistry, RWTH Aachen, May 10 and 11, 2023. Program and a report, NFDI4Chem, Jun. 29, 2023; Clark, A.M., (Oct 18, 2020) Coordination InChI for inorganics: now with stereochemistry. Cheminformatics 2.0.; IUPAC, 2019. Project No.:2019-017-2-800 Enhanced recognition and encoding of stereoconfiguration by InChI tools. Chair Andrey Yerin, Start Date: 9 Sep 2019; Yerin, A. (Aug. 16, 2017). Improvements in InChI treatment of stereoconfiguration.(16 p.) @ Status and Future of the IUPAC InChI: context and use cases, NIH, Bethesda, MD, Aug. 16-18, 2017. Cf. Apodaca, R.L. (February 9th 2022). V3000 Molfile Enhanced Stereochemistry Representation. Depth-First. (and refs. herein)

CINF Fresh Takes and the Super Genius: Sharing News From the Field, and Envisioning the Future of Chemical Information 8/25 PM

4524140 - Role of Abstract Sifter, a chemistry-enabled Excel tool, in the chemical data ecosystem. Nancy Baker, ParlezChem

We demonstrate how the Abstract Sifter, originally a literature tool, uses API calls to PubChem, DSSTox, and CAS to provide a rich menu of functions that can retrieve chemical data structure diagrams, SMILES, chemical properties and assay results. We also show how chemical entities can be mined from biomedical and chemistry literature can be combined with chemical data to build robust chemical datasets for further analysis and testing.

Add. info: Abstract Sifter (Free download Abstract Sifter v.8 & v.9, include an user guide); Baker NC, Knudsen TB and Williams AJ. (June 6, 2026) Abstract Sifter version 8: Focus on the chemical literature [v 2]F1000Research 2026, 14(Chem Inf Sci):260 (https://doi.org/10.12688/f1000research.160617.2)

Figure 1. Abstract Sifter version 8 introduces a new EPA Custom ribbon with new functionality specific for working with chemicals and chemical literature

Note: Abstract Sifter v.9 is planned to be extended to query SureCHEMBL patents:

CINF Machine Learning and AI for Organic Chemistry 8/25 PM

4530332 - Chemist's perspective on publicly available AI retrosynthesis models Tom Nakotte, Lawrence Livermore National Laboratory

We report a chemist-centered comparison of four retrosynthesis platforms, IBM RXN, ASKCOS, Molecules[Uses AstraZeneca open-source AIZynthFinder – ab], and FLASK (Foundation Learning AI for Synthesis Knowledge, an internally developed LLNL platform). Fifteen targets were selected to span both familiar transformations and less common edge cases.

Template-driven frameworks [ASKCOS, AIZynthFinder] were most reliable when a transformation closely matched curated precedents, but they frequently struggled when the target lay outside template coverage or required nonstandard representations, leading to missing routes or chemically incoherent suggestions. In contrast, neural approaches [IBM RXN, FLASK] generalized more flexibly across common reaction classes and often proposed step-economical disconnections, but outputs sometimes lacked essential experimental details (reagents, stoichiometry, or conditions), limiting direct actionability at the bench.

Add. info: Nakotte, T.(LLNL), 11 Sept 2025. Comparative Analysis of Template-Based and Neural Network Approaches for Computer-Aided Retrosynthesis: ASKCOS vs. IBM RXN. Submitted to the Agents4Science 2025,Oct. 22, 2025. Submission No. 107, 10 p. See selected bibliography on retrosynthesis platforms referenced in the presentation in Footnote 2 below.

CINF Machine Learning and AI for Organic Chemistry 8/25 PM

4530332 - Chemist's perspective on publicly available AI retrosynthesis models Tom Nakotte, Lawrence Livermore National Laboratory

Wed. Aug. 26, 2026

CINF Analyzing and Dealing with Diverse Data and Linked DataSets in New and Different Ways for ML Model Development 8/26 AM

4544674 - Open-source quantitative reaction condition dataset extracted from USPTO. Xiaoqi Sun, and Connor Coley, Massachusetts Institute of Technology

To address the gap of a large-scale, open-source dataset tailored to comprehensive reaction condition we construct a procedure-aware condition dataset from USPTO. By adapting recent information extraction workflows, we convert unstructured procedural text into structured Open Reaction Database (ORD) records that capture both reaction-level information such as inputs, conditions, and workup operations, and component-level details including compound identities and quantities. We present the dataset and discuss its applications including condition recommendation, reaction condition space analysis, and benchmarking existing data resources.

ANAL / CINF Analyzing and Dealing with Diverse Data and Linked DataSets in New and Different Ways for ML Model Development 8/26 PM

4541545 - Advances in reaction extraction and wrangling Rachael Pirie, John Mayfield, and Roger Sayle, NextMove Software Limited

NextMove Software’s Pistachio dataset consists of ~22 million reactions extracted automatically from text and sketches in patent documents, and as such grows not just from new patents but with improvements to the extraction pipeline, which increases the yield from historic data too. Recent improvements include: unified (instead of independent) extractions reactions reported in text and sketches; capture synthesis of related compounds by similar methods “summarised in patents as a single example described in full, and a set of alternative reagents or products provided that follow the same procedure”, better handling of chiral separation reactions using enhanced CXSMILES to increase NameRXN coverage, etc.

Add. Info: Pistachio [Version 2026-06-30 (2026Q2)] Reaction Data, Querying and Analytics (includes technical demonstration of the reaction search and browser, video, 2:53 Apr 2022); Mayfield, J., Lagerstedt, I., Sayle, R., May 19, 2021. How reaction data, including conditions and operations, are captured, represented, managed and analyzed in Pistachio. NIH Virtual Workshop on Reaction Informatics, May 2021 ( See also Dr. Wendy Warr’s workshop report (2021), pp. 34-36); Mayfield, J., March 24, 2021. 13,118,970 Reactions and Counting. NextMove Software Blog.; Mayfield, J., Dec.11, 2017. Pistachio: Search and Faceting of Large Reaction Databases. NextMove Software. (with a link to the presentation (30 p.) at the ACS Fall 2017, Washington, DC, Aug. 20-24, 2017)

MEDI Poster Session, 8/26, 6 PM CT

4538097 - Navigating the pharmaceutical innovation landscape: A comprehensive patent analysis of global therapeutic development (2020-2025) | Poster Board #1145 Qiongqiong Zhou, Kavita Iyer, Janet Sasso, and Julian Ivanov, CAS, Trupti Thite, Preeit Jain, Ankush Maind, and Umesh Balande, ACS International India Pvt. Ltd

This study leverages the CAS Content Collection™, analyzing more than 365,000 pharmaceutical patent families (2020-2025) encompassing diagnostics and therapeutics using Proprietary CAS indexing methodology applied by expert scientists to chemical structures, target proteins, therapeutic concepts and disease ontology. Our framework examined: … therapeutic modality trends spanning small molecules, biologics (antibodies, peptides), RNA therapeutics, cell/gene therapies, advanced drug delivery systems, radiopharmaceuticals, and microbiome therapeutics which demonstrated unprecedented diversification. Granular CAS indexing and advanced AI-driven analytics enabled to identify dominant targets with intensive activity and emerging targets with accelerating filing rates. Cross-referencing targets with disease areas and modalities revealed white space opportunities and overcrowded competitive zones.

Add. Info: Add, Info: ACS Webinar Patent Signals: Mapping the Future of Pharma Innovation (Jul. 15, 2026, video, and Slides, 30 p. w/ Janet Sasso, Kavita Iyer and and Qiongqiong Angela Zhou, CAS); Iyer, K.A., Zhou, Q., Thite, T., Jain, P., Sasso, J., Ivanov, J., Maind, A., Jun 2026. Global Pharmaceutical Patent Intelligence: Connecting the dots between technology, targets, and therapies (CAS Insights Report). CAS., 189 p.(Sec. 2 Methodology, p.11-12; Sec. 3 Patent landscape, p.12-29; CAS TrendScape map, listed in the List of Figures [p.183])]; Executive Summary.; Dr Zhou LinkedIn post (Jun 15 2026)

COMSCI Cosmetic Technology Showcase: Industry Insights 8/26 PM

4569599 - From innovation to protection: Strategic patent approaches in the cosmetics industry Mareesa Frederick, Finnegan Henderson Farabow Garrett and Dunner LLP

A thoughtful patent strategy ….enables companies to protect novel active ingredients, formulations, delivery systems, manufacturing processes, and other technological advancements while establishing a competitive market position. Effective navigation of the patent pathway begins with assessing patentability and FTO…. A strategic filing approach that aligns with target markets and business objectives, combined with ongoing monitoring of competitor patent activity and the pursuit of follow-on innovations, can create a strong and layered patent portfolio.

Add. info: Silverman, L., Reaney, S., Sanders, M. (Morrison Foerster) Oct. 28, 2025. Patent Protection Is the Foundation of the Future for Cosmetics. Bloomberg Law; Evenson, J., Beynon, N., Hickling, M., Feb. 2025. Cosmetics Patents. CMS. 14 p. (trends in cosmetic patent filings, 2013-2023; IPC A61K 8/00 & A61Q; Patseer); Aquilina, J., Stewart, G, Apr. 17, 2024. Protecting the Product: Beauty Products. Quarles (design patents & trade dress); Pazik, A., Gągała, U. Nov. 24, 2023. Patent trends in cosmetics industry for 2024. JWP.; Pellow, K., April 27, 2023, Cosmetics: Getting under the skin of patent protection in this…. Abel + Imray.; Grosshuesch, K., Valenti, A.D., Birbach, N., Sept. 20, 2022. 3 Reasons Beauty Cos. Should Consider Pursuing Patents. Goodwin.; Zeng, A.D. Nov. 30, 2020.Utility Patents: Another Strategy to Protect Your Beauty and Cosmetic Portfolio. Knobbe Martens.; Sen, S., Bhattacharya, N., 2025. Innovating Beauty: Unveiling the Role of Patents in the Cosmetic Industry. Journal of Intellectual Property Rights (JIPR) 30(3), 383–396. https://doi.org/10.56042/jipr.v30i3.10973 (patenting trends, 2003-2023, Questel Orbit)

Thu., Aug. 27, 2026

CINF From Insight to Impact: Frameworks for Reliability and Reproducibility for Agentic AI in the Life Sciences 8/27 AM

4538800 - Unlocking biochemical reactions from full text a hybrid LLM-ontology approach to combat hallucination. Shadrack Jabes Barnabas, Claudia Bobach, Matthias Irmer, Christoph Ruttkies, and Arya Anil, Digital Science & Research Solutions Ltd

This work presents a hybrid framework that integrates LLMs with formal chemical ontologies and atom-mapping validation to automate the extraction of enzymatic reactions from full-text sources, including PMC Open Access articles and patents. (1) a document annotation with Digital Science (DS) [formerly OntoChem] ontologies; (2) identification of paragraphs containing complex enzymatic descriptions with 93% accuracy classifier (3) identification of specific reaction roles, including substrates, products, and modifiers with an LLM-based extractor (4) resolution these machine-generated entities against the DS ontologies, ensuring named entities are mapped to valid chemical structures. For validation, Atom-to-Atom Mapping (AAM) is used to verify the presence of a reaction center, for complete reactions where both reactants and products are present.

=====================

Footnote 1. Examples of RefChem compounds linked to patents in Pubchem (related to the presentation 4538125 - Improving the chemical entity integration in PubChem: introducing RefChem. Evan Bolton, et al, NCBI (CINF Skolnik Award Symposium: Antony Williams, 8/25, PM)

Examples: (1) RefChem:1092876, CID 73995022, Npc212957, Casein from bovine milk (Kappa casein), C₈₁H₁₂₅N₂₂O₃₉P, Linked to 9,848 patents from Patentscope through InChI Key BECPQYXYKAMYBN-UHFFFAOYSA-N; (2) Search for patents from Patentscope for compounds with CID from RefChem collection: ("PATENTSCOPE (WIPO)" [SRC]) AND "RefChem"[Synonym] => 282963 CIDs, one of them, CID 10137, C₁₅H₂₄, RefChem:1093009, Npc279404, PDSNLYSELAIEBU-PQNSMMMDSA-N

US20200088698 - Method for determining whether or not infection with bursaphelenchus xylophilus, where the compound is listed in the description as Junipene, or (+)-longifolene, CAS 475-20-7, CID 1796220 (none of this chemical names are in RefChem) (The same InChI Key, PDSNLYSELAIEBU-GUIRCDHDSA-N, but different, more defined stereochemistry:

(3) AGRO 100; RefChem:553192; Nucleic acid aptamer binding to tumor cell nucleon…. Compound CID:121494112, MF:C₂₆₀H₃₂₂N₁₀₃O₁₆₃P₂₅, MW:8272.0 g/mol, CAS 301636-59-9, Sequence GGTGGTGGTGGTTGTGGTGGTGGTGG, DFYPFJSPLUVPFJ-QJEDTDQSSA-N.

It linked to 195 patent documents, Starting with US-2008119432-A1 with the earliest priority date 2004-10-19, and published May 2008, The compound is listed on p.3, par. 103 of the description under its name AGRO100, a G-rich oligonucleotide, as the first modified-oligonucleotide drug, and listed in machine extracted concepts (Chemical compounds) by its full IUPAC name ([(2R,3S,5R)-5-(2-amino-6-oxo-1H-purin-9-yl)…]. Note that CID:121494112 is linked in Pubmed for the first published disclosure to Girvan, A.C. Mol Cancer Ther 2006 Jul;5(7):1790, doi: 10.1158/1535-7163.MCT-05-0361. where for AGRO100, a sequence is given as 5′-d(GGTGGTGGTGGTTGTGGTGGTGGTGG)-3′ on p.1791)

Footnote 2. See selected bibliography on retrosynthesis platforms referenced in the presentation 4530332 - Chemist's perspective on publicly available AI retrosynthesis models Tom Nakotte, Lawrence Livermore National Laboratory (CINF Machine Learning and AI for Organic Chemistry 8/25 PM)

(1) IBM RXN for Chemistry, https://rxn.app.accelerate.science/rxn/sign-in (accessed 8.27.26, reaching end of service on October 28, 2026); Lopes, Helder (helderlopes97), Aug. 6, 2026. rxn4chemistry/rxn-sandbox. https://github.com/rxn4chemistry/rxn-sandbox; Schwaller, P., Hoover, B., Reymond, J.-L., Strobelt, H., Laino, T., Apr. 7, 2021. Extraction of organic chemistry grammar from unsupervised learning of chemical reactions. Science Advances 7(15), eabe4166. https://doi.org/10.1126/sciadv.abe4166; IBM Research, Feb. 9, 2021. Thieme trains IBM RXN for Chemistry with high-quality data.; Schwaller, P., Laino, T., Gaudin, T., Bolgar, P., Hunter, C.A., Bekas, C., Lee, A.A., 2019. Molecular Transformer: A Model for Uncertainty-Calibrated Chemical Reaction Prediction. ACS Central Science 5, 1572–1583. https://doi.org/10.1021/acscentsci.9b00576; (2) ASKCOS MIT, https://askcos.mit.edu; ASKCOS Documentation Computer-aided tools for Organic Synthesis; Tu, Z., …. Coley, C.W., 2025. ASKCOS: Open-Source, Data-Driven Synthesis Planning. Acc. Chem. Res. 58(11), 1764–1775. https://doi.org/10.1021/acs.accounts.5c00155 (template-based and template-free strategies for retrosynthetic planning); (3) Ramos, M.C., Cox, S., Dec. 4, 2025. Molecules: An AI Agent for Chemistry. Edison Scientific.; (3a) Genheden, S., et al.., 2020. AiZynthFinder: a fast, robust and flexible open-source software for retrosynthetic planning. J Cheminform 12(1), 70. https://doi.org/10.1186/s13321-020-00472-1 (a neural network trained on a library of known reaction templates); Genheden, S., et al. Dec. 2020. MolecularAI /aizynthfinder. https://github.com/MolecularAI/aizynthfinder; Genheden, S., Jan 2022. AiZynthFinder (1:01:47); (4) van Essen, Brian., Apr 8, 2025. FLASK: Foundation Learning AI for Synthesis Knowledge. 16th Annual HPC-AI Swiss Conference, April 7 - 9, 2025, Lugano, Switzerland (Video, 32:50); Jaffe, Elliot, Feb, 2 2026. FLASK Copilot offers scientists new chemical recipes. LLNL Data Science Institute ; Thomas, J., May 18, 2026. LLNL showcases AI-enabled science, national security and energy innovation at AI+ Expo. Lawrence Livermore National Laboratory.

Search This Blog

Patent & Patent Information Events

Highlights of ACS Fall 2026, August 23-27, 2026 Part 2 Tue. Aug. 25 2026 - Thu, Aug. 27, 2026

Comments

Post a Comment

Popular posts from this blog

Patent and Patent Information Events in January 2026 (Updated 1/27/2026)

Patent and Patent Information Events in December 2025 (Updated 12/21/2025)

Patent and Patent Information Events in June 2026 (Updated 6/23/2026)