AI-augmented innovation: Enabling DMTA with digital-physical synergy

Andrew Anderson, Vice President, Innovation and Informatics Strategy at ACD/Labs, looks at how AI is enabling true drug design. Innovators in the pharma and biopharma sectors are actively seeking systems that accelerate the discovery, development, and commercialisation of new products. Of particular interest are systems to support the implementation of continuous and mutually-reinforcing digital-physical feedback loops. Here, digital tools and methods are used to enhance physical processes, and feedback from these improved physical processes informs progressive digital advancements. Each transition within the non-digitalised design-make-test-analyse (DMTA) cycle (from Design, to Make; from Make to Test, etc) often requires significant human transposition and translation of information from systems to bridge the gap between different stages and domains of expertise. However, ineffective transition management in DMTA can lead to: Productivity loss—translating from one domain to the next often requires practitioners to leverage subject matter experts to map important, domain-specific terms (from design systems, to execution systems, to analysis systems). Transposition/transcription errors—incorrect manual transposition and/or transcription of numbers and text into software interfaces can result in failed experiments, faulty analyses, or incorrect decisions. Put simply, avoiding these risks across the various DMTA transitions and from digital-to-physical steps allows for a virtuous DMTA cycle. Digital-physical DMTA synergies in drug discovery A variety of specific DMTA cycles1 are required for successful candidate nomination in drug discovery organisations. The following summarises the dual-purpose of DMTA cycles during lead optimisation in drug discovery. AI-enabled drug design—the design step The design phase of the DMTA cycle in drug discovery addresses two key questions: What to make? Specifically, what composition of matter is most suitable for modulating the applicable drug target’s biochemical function? Virtuous approaches involve the use of generative AI tools to produce a structure activity relationships (SAR) map. Once sufficient SAR information has been obtained for a lead series, medicinal chemists attempt to optimise for potency, selectivity, and overall druggability. Modern drug discovery organisations have implemented a range of generative AI4,5,6 that can produce an optimised set of target compounds. How to make it? In order to confirm the success of sequential rounds of generative AI, medicinal chemists must design efficient synthetic routes for each of the target compounds; and also minimise the total number of reactions required to produce confirmatory test articles for all AI-recommended target compounds. Efficient approaches involve the use of well-trained retrosynthesis tools7 medicinal chemists use output to map to experiment execution systems, including to applications8 which generate machine-friendly master procedure lists. These instructions include material dispenses, reaction operations, and sample preparations—enabling efficient preparation and execution of 'Make' tasks required for preparation of test articles. AI-enabled synthesis execution in drug discovery—the make step Upon completed design, medicinal chemists can execute each unit operation, from material dispensing, through reaction initiation, to final workup and sample preparation (Figure 2). In modern synthesis labs, the format of machine instructions must conform to the format requirements of each instrument supporting each unit operation. The output results in a set of test articles, labelled with appropriate machine-and-human-legible material identifiers. These identifiers serve to associate test article information stored in appropriate software applications including output material identity, material metadata, container or vessel IDs. AI-enabled testing in drug discovery—the test step Similarly to design, the testing of output materials (i.e, purified target compounds) serves a dual purpose: 1. Subjecting materials to a wide range of project-applicable bioassays to confirm the performance of the target compounds as possible (future) drugs. 2. Identity and quantitative compositional testing to assure accurate SAR9. As such, the modern DMTA cycle accounts for all applicable tests required to inform the overall lead optimisation effort. Samples of output materials from reaction execution steps are prepared, then labelled with system-derived sample identifiers, including: Reaction quality control samples—sample type (preprocess, in-process, and post process), sample preparation information. Product registration samples—including chemical identity, purity, salt stoichiometry, and any applicable physical form information that is required for assessing bioassay results for SAR map. Test article samples—including sample concentration, assay role (e.g., control, standard, blank, replicate number, etc). Predicted properties—intended to complement the measured bioassay and QC results, enhancing the utility of the overall SAR Model. AI-enabled analysis in drug discovery—the analyse step The first analysis activities support the processing and interpretation of product QC and bioassay data. The variety of tools that support this primary analysis step has been described elsewhere.10,11,12 Modern drug discovery organisations intend to leverage the output of these analyses, by: 1. Aggregating processed data into a data warehouse 2. Implementing a rigorously enforced controlled vocabulary for all bioassay and product QC Results, as well as associated metadata structured as JSON Objects 3. Relating the test results by applicable sample, corporate identifier, and in-silico properties (generated for product registration). Scientists can then update applicable SAR maps (Figure 1) based on the bioassay test results. Value realisation of digitalised DMTA Enabling the modern, virtuous DMTA cycle offers significant benefits to drug discovery productivity. This approach allows for the generation of new target compounds and facilitates a seamless and iterative process. Consequently, the integration of structured data, controlled vocabulary, and advanced analysis tools not only enhances the efficiency of the drug discovery process but also accelerates the development of innovative therapeutics. Ultimately, the virtuous DMTA cycle allows project teams to optimise lead series faster, arriving at molecules which meet clinical candidacy specifications. From DDW Volume 26 – Issue 3, Summer 2025 Read the digital issue References: DMTA Cycles include disease pathology research, biological target and pathway analysis, molecular modality selection, in-vitro assay development, in-vivo study design, and test article scale-up research. Certainly, all require effective DMTA cycles to matriculate through to clinical studies. Frey, K. (2020). Structure activity relationship (SAR) maps. Curr. Pharm. Teach. Learn., 12(3), 339-346. https://doi.org/10.1016/j.cptl.2019.12.014 Vangala, S.R., et al. (2024). Suitability of large language models for extraction of high-quality chemical reaction dataset from patent literature. J. Cheminform., 16, 131. https://doi.org/10.1186/s13321-024-00928-8 S. Reddy N, et al., "Leveraging Latent Evolutionary Optimization for Targeted Molecule Generation," 2024 IEEE Congress on Evolutionary Computation (CEC), Yokohama, Japan, 2024, pp. 1-8. https://ieeexplore.ieee.org/document/10611790 Ochiai, T., et al. (2023). Variational autoencoder-based chemical latent space for large molecular structures with 3D complexity. Commun Chem, 6, 249. https://www.nature.com/articles/s42004-023-01054-6 M. C. Ramos, et al. (2024, June 26). A Review of Large Language Models and Autonomous Agents in Chemistry. arXiv. https://arxiv.org/html/2407.01603v1 Discovery at Your Fingertips | SYNTHIA™ Retrosynthesis Software. (2025, April 16). Synthia Online. https://www.synthiaonline.com/ Software to support HT experimentation workflows. Katalyst D2D, version 2024, Advanced Chemistry Development, Inc. (ACD/Labs), Toronto, ON, Canada. www.acdlabs.com/katalystd2d It should be noted that organizations will also subject output materials to such analyses in defense of composition of matter claims in IP Applications. Basics of claim drafting for utility patent applications. (Retrieved 2025, Apr. 22). USPTO. https://www.uspto.gov/sites/default/files/documents/InventionCon2021WhatsinaPatent ClaimWorkshopFinalstakeholders.pdf Mercer, K. (2023, Oct. 24). Enabling End-User HTE at GSK . ACD/Labs Driving Efficiency with Spectrus Symposium. https://www.acdlabs.com/resource/enabling-end-user-hte-at-gsk/ Standardization of Analytical Data: Best Practices. (2025, Jan. 6). ACD/Labs. https://www.acdlabs.com/resource/standardization-of-analytical-data-best-practices/ FAIR Annotation of Bioassay Metadata. (Retrieved 2025, Apr. 22). Pistoia Alliance. https://fairtoolkit.pistoiaalliance.org/use-cases/fair-annotation-of-bioassay-metadata/ About the author: Andrew Anderson leads product and technology management efforts as VP of Innovation and Informatics Strategy at ACD/Labs. He previously worked in Technology Scouting at PepsiCo. Anderson earned a B.Sc in Chemistry and MBA from San Diego State University and started his professional career as an Analytical Chemist and Regulatory CMC Specialist at Pfizer.