Services to Support AI-Informed Drug Discovery & Development

The integration of artificial intelligence (AI) into drug discovery and development pipelines has the potential to revolutionize the field, enabling effective medications to be developed more efficiently and effectively than ever before. This technology has been particularly exciting for early discovery workflows, enabling large data sets to be synthesized and analyzed for the identification of relevant disease targets and biomarkers, the design of promising small-molecule and biomolecular binders, and the prediction of pharmacological properties. However, despite these promising opportunities, scientists should remain aware of the limitations of AI-powered approaches and thoughtfully balance their use with the use of biological and biochemical assays. By designing a research workflow that incorporates both AI-powered and traditional methodologies, researchers can maximize experimental utility and accelerate the drug discovery and development process. Below, we’ll highlight key preclinical research stages where AI has transformative potential, as well as others where wet-lab approaches reign supreme.

Strength: Molecule generation

AI and machine learning approaches have shown significant potential to accelerate the initial stages of drug discovery through both the de novo design of novel molecules (AI-designed or AI-assisted compounds) and the selection of promising molecules from existing repertoires (AI-repurposed compounds). These methods leverage structural models and other large biological and chemical data sets to identify compounds that seem likely to bind a target protein or have a particular functional effect. In addition to small molecule generation, such approaches have shown significant promise in the design of novel antibodies and in the development of cell and gene therapies. The success of AI-based workflows for molecule generation relies upon access to reliable input data, which can be used to train AI models and synthesized to obtain compound recommendations. Momentum’s data products – the Protein Turnover Atlas™, Mouse Turnover Atlas™, and Proteome Atlas™ – are well-suited for this purpose, as they contain clinically relevant, high-quality data that has been curated by scientific experts.

Strength: Prediction of toxicity, interactions, and other key pharmacological parameters

AI models have proven well-suited for the development of PK-PD models, prediction of drug-drug interactions, and estimation of drug toxicity. Such efforts offer the promise of significantly acceleration drug development timelines, allowing key pharmacological parameters to be forecast well in advance of clinical trials. These approaches leverage vast amounts of existing information regarding drug structure and activity to make informed predictions about the likely behavior of candidate compounds. To this end, researchers have developed powerful AI platforms designed to predict the ADMET (Absorption, Distribution, Metabolism, Excretion, and Toxicity) properties of a given drug, as well as to predict potential off-target interactions and their safety effects. Armed with this data, drug developers can prioritize drug candidates for further exploration, optimize the pharmacokinetic and pharmacodynamic properties of compounds of interest, and design additional experiments to directly investigate phenomena predicted by AI analysis.

Limitation: Biochemical and functional validation

While AI approaches can readily generate and prioritize molecules for a particular application, the ultimate validation of these compounds requires experimental interrogation. This can be accomplished by biochemical means (such as ASMS) or by functional screening (as by RapidFire-MS). Target identification and deconvolution technologies – including CHIPP and SPICE – can also provide valuable confirmation of a drug:target interaction. These ‘wet lab’ approaches are the gold-standard for validating a true physical and functional relationship between the compound of interest and a desired target and remain a critical component of any AI-enabled drug discovery workflow. As researchers Xiaomeng Liu and Huanxiang Liu eloquently summarize in a recent paper, “AI-generated molecules should ultimately be treated as testable hypotheses rather than conclusions.”

With respect to pharmacological parameters (bioavailability, toxicity, PK-PD, etc.), in cellulo and in vivo experiments remain necessary to evaluate AI predictions in a physiological context. Due to the immense complexity of biological systems, AI approaches may fail to capture the full repertoire of drug responses, especially when trained on limited data sets. For this reason, wet-lab experiments will continue to be crucial in the evaluation of compounds that have been generated, prioritized, or otherwise informed by AI tools and platforms, ensuring that the impressive data-synthesis capacity of AI is balanced by downstream follow-up analysis at the bench.

What services does Momentum offer to support AI-informed drug discovery?

Momentum’s data products – including the Protein Turnover Atlas™, Mouse Turnover Atlas™, and Proteome Atlas™ – can serve as valuable sources of high-quality data for the development, training, and refinement of AI models. By leveraging these curated databases, researchers can ensure that their AI-based models and platforms are based on biologically relevant data from a reputable source. Furthermore, virtually all of Momentum’s services can be harnessed to verify predictions or hypotheses generated via AI workflows. Of particular note, ASMS and covalent binding assays can be used for the biochemical validation of small molecule hits, while our suite of target deconvolution technologies can provide clear confirmation of drug:target binding. Assays such as ProteomeScout™, TurnoverScout™, and QuantScout™ can verify predicted effects on protein abundance and turnover, further supporting the development of novel therapeutics.

To learn more about Momentum’s services to support AI-informed drug discovery and development, connect with our team.

 

Sources

Auclair JR, Joung J, Singh MA, Debauve G & Singh R. Artificial Intelligence and the Transformation of Cell and Gene Therapy Development. Pharmaceutics 18, 356 (2026).

Cortial L, Montero V, Tourlet S, Del Bano J & Blin O. Artificial intelligence in drug repurposing for rare diseases: a mini-review. Front Med 11, 1404338 (2024).

Huang W, Wang X, Chen Y, Yu C & Zhang S. Advancing drug-drug interactions research: integrating AI-powered prediction, vulnerable populations, and regulatory insights. Front Pharmacol 16, 1618701 (2025).

Huang Z, Denti P, Mistry H & Kloprogge F. Machine Learning and Artificial Intelligence in PK-PD Modeling: Fad, Friend, or Foe? Clin Pharmacol Ther 115, 652-654 (2024).

Khairil L, Benny KHS, Jerry J, Khatib FM, Ramli MDC & Kumar S. AI in Drug Discovery: Clinical Failures, Regulatory Reality, and the Validation Crisis Behind the Hype. Pharmaceuticals 19, 916 (2026).

Liu X & Liu H. Evaluating AI-Generated Molecules for Drug Discovery: From Generic Metrics to Translational Readiness. Int J Mol Sci 27, 5916 (2026).

Niazi SK. Artificial Intelligence in Small-Molecule Drug Discovery: A Critical Review of Methods, Applications, and Read-World Outcomes. Pharmaceuticals (Basel) 18, 1271 (2025).

Pathan I, Raza A, Sahu A, Joshi M, Sahu Y, Patil Y, Raza MA & Ajazuddin. Revolutionizing pharmacology: AI-powered approaches in molecular modeling and ADMET prediction. Med in Drug Disc 28, 100223 (2025).

Seyhan AA & Carini C. Artificial Intelligence in drug discovery and development: current landscape, challenges, and future perspectives. Crit Rev Oncol/Hematol 226, 105461 (2026).

Sumita M, Ishida S, Yoshizoe K, Tamura R, Terayama K & Tsuda K. Molecular Design with Artificial Intelligence: Progress and Perspectives for Small Molecules. Chem Rev 126, 3007-3054 (2026).

Vecchietti LF, Wijaya BN, Armanuly A, Hangeldiyev B, Jung H, Lee S, Cha M & Kim HM. Artificial intelligence-driven computational methods for antibody design and optimization. MAbs 17, 2528902 (2025).

Zhang R, Wen H, Lin Z, Li B & Zhou X. Artificial Intelligence-Driven Drug Toxicity Prediction: Advances, Challenges, and Future Directions. Toxics 13, 525 (2025).