AI in Pharmacy and Pharmaceutical Sciences

AI and Machine Learning are Integrated Across All Areas of Research and Discovery

AI and machine learning are no longer confined to a single lab in the School of Pharmacy — they are the connective layer underneath drug discovery, genomics, formulation science, toxicology, pharmacogenomics, and clinical decision support. This area gathers the faculty who are building and applying those models, from generative chemistry and protein structure prediction to deep-learning tools that flag drug-induced kidney injury at the bedside.

Their work spans the full arc of pharmaceutical science: reading DNA's regulatory grammar, designing molecules and enzymes computationally, predicting how a drug will behave in the body, and deploying clinical decision-support tools that put those predictions in front of pharmacists and physicians in real time.

Faculty Leading This Work

Select a name to see how each lab is using AI and machine learning, along with representative publications, active grants, and trainees.

Sravan Kumar Patel, PhD
- Drug Delivery & Formulation

Uses deep learning to untangle the non-linear relationships between formulation parameters and film quality, to track nanoparticles in biological fluids, and to make sense of noisy spectral data from hand-held drug-analysis devices.

AI/ML Focus

• AI/ML models of the multifactorial relationship between vaginal-film formulation and processing parameters and resulting film quality.
• Deep learning for tracking nanoparticle diffusion in biological matrices (vaginal and lymphatic fluids) from confocal video, including synthetic-data generation to boost predictive power.
• Deep-learning chemometrics for denoising hand-held NIR spectrophotometer data for rapid drug analysis in biological fluids.

Representative Publications

• ANN models assessing the impact of hot-melt extrusion process parameters on film quality. DOI: 10.1016/j.ijpharm.2019.118715
• Perspective review with the AAPS Nanotechnology community on AI/ML methods in drug discovery and development. DOI: 10.1007/s11095-024-03798-9

Trainees

• Helena Zhang — vaginal film optimization
• Parthiv Reddy Bandi (Undergrad) — nanoparticle tracking

Junmei Wang, PhD
 - Computational Drug Design

Applies machine learning and deep learning to expand druggable chemical space, predict biomolecular structure and binding affinity, and rationally design biased ligands.

AI/ML Focus

• Machine-learning interatomic potentials (MLIPs) combined with molecular mechanics to model enzyme mechanism, stereoselectivity, and catalysis.
• Generative adversarial networks for de novo molecule generation in low-data, drug-like chemical spaces.
• Deep learning applied to protein–ligand and protein–peptide structure and binding-affinity prediction, including conditional-diffusion extensions to AlphaFold3.

Representative Publications

• Conditional diffusion built on AlphaFold3 for accurate site-specific co-folding. PNAS, 2025.
• MAPLE, an automated MLIP-based platform for reaction modeling and enzyme design. Chemical Science, 2026.
• Multiscale MLIP/molecular-mechanics framework for Diels-Alderase catalysis mechanism and stereoselectivity. Nature Communications, 2026.
• 11 total AI/ML publications in the last two years, including in J Chem Inf Model, Brief Bioinform, and J Chem Theory Comput.

Active Grants

• NIH R01 GM149705 (2023–2027), PI — AI-powered biased ligand design using interaction-profile scoring functions and Drug-GAN de novo design.
• NIH R35 GM163906 (2026–2031), PI — Next-generation AMBER force field (GAFF3) for AI-driven computer-aided drug design.

Trainees

• Taoyu Niu
• Yue Liu
• Xujian Wang
• Yixuan Hao
• Lianjin Cai
• Jingchen Zhai (Postdoc)

Min Zhang, PhD
 - Computational Toxicology & Pharmacogenomics

Integrates large-scale multi-omics and single-cell/Perturb-seq data with AI/ML to identify biomarkers, characterize treatment resistance, and predict drug toxicity.

AI/ML Focus

• Elastic-net and other ML models integrating multi-dimensional genomic and drug-response data to identify predictive biomarkers.
• INSIGHT, a computational framework for in silico screening of drug-induced hepatocellular toxicity from transcriptomic data.
• Single-cell RNA-seq and Perturb-seq analysis of cellular heterogeneity and gene-regulatory response to genetic and therapeutic perturbation.

Representative Publications

• INSIGHT, a computational framework for in silico prediction of drug-induced liver injury. Toxicological Sciences, 2024 — corresponding author.
• Single-cell analysis showing lncRNA EPIC1 suppresses type-I interferon signaling as a target to enhance TNBC response to PD-1 inhibition. Science Signaling, 2025 — co-corresponding author.
• Machine-learning models integrating genomic and drug-response data across 1,005 cell lines and 265 compounds. Nature Communications, 2018 — co-corresponding author.
• Machine-learning classifiers using rat toxicogenomic data to predict human drug-induced liver injury (83% accuracy, leave-one-compound-out CV). Chemical Research in Toxicology, 2012 — first author.

Trainees

• Saisai Ma
• Yaohua Ni
• Yueshao Zhao
• Yue Wang

Qihao Wu, PhD - Metabolomics & Host–Microbiome Chemistry

Uses AI/ML to accelerate small-molecule discovery at the host–microbiome interface and to scale up untargeted metabolomics across tens of millions of mass spectra.

AI/ML Focus

• AI- and ML-enabled organization and analysis of more than 75 million public tandem mass spectra to map uncharacterized diet–microbiome chemistry.
• AI-assisted metabolomics and computational modeling to identify enzyme–metabolite relationships at the host–microbiome interface.

Publications in Preparation

• Diet–microbiome chemical space atlas — PI; systematic mapping and prioritization of previously uncharacterized microbial metabolite families.
• Host–microbiome small-molecule metabolism project — PI; AI-assisted identification of enzyme–metabolite relationships shaping small-molecule metabolism.

Trainees

• Wenchao Yu, PhD — chemical space atlas
• Yan-Song Ye, PhD — enzyme discovery

Da Yang, MD, PhD
 - Genomics & Precision Oncology

Builds AI that reads and ultimately redesigns the regulatory "grammar" of the genome — combining CRISPR screening and single-cell data with generative models to predict and rewrite disease programs.

AI/ML Focus

• Causal AI trained on CRISPR/Perturb-seq data rather than observational data alone, to learn how genomic sequence, chromatin, and cell state interact.
• Predictive pharmacogenomics integrating genomic, transcriptomic, and drug-response data to forecast chemotherapy, targeted-therapy, and immunotherapy response.
• Variational autoencoders for generative modeling of cell states, and early work toward a "Cell GPT" that predicts cell response and proposes interventions from a genomic or drug prompt.

Representative Publications

• Elastic-net models linking non-coding pharmacogenomic features to drug response across 1,005 cancer cell lines and 265 drugs. Nature Communications, 2018 — senior/corresponding author.
• Genome-wide CRISPR activation screening combined with immunogenomics to map tumor immune-evasion regulators. Science Advances, 2022 — senior/corresponding author.
• A computational "shift-ability" algorithm applied to 41,000+ compound perturbations to prioritize drug combinations that resensitize anti-PD-1-resistant tumors. Nature Communications, 2024 — senior author & PI.
• 11 AI/ML-related publications total, spanning 2011–2026, including work in Cancer Cell, JAMA, and Cell Genomics.

Active Grants

• NIH/NCI R01CA282704 (2023–2028), PI — Enhancer RNAs and MYC–chromatin interaction; a foundation for AI models of genome grammar.
• NIH/NCI R01CA272866 (2023–2028), PI — LncRNA EPIC1 and immunotherapy resistance in breast cancer.
• NIH/NCI R01CA255196 (2020–2026), PI — LncRNA-responsive regulation of drug metabolism and disposition.

Trainees

• Yueshan Zhao
• Yu Zhang
• Sihan Li
• Yifei Wang
• Yue Wang
• Zehua Wang
• Dhamotharan Pattarayan
• Weiwei Guo
• Xiaofei Wang

LiRong Wang, PhD
 - Chemogenomics & Clinical AI

Develops deep-learning tools that mine electronic health records to predict substance-use, psychiatric, and adverse-event outcomes, alongside chemogenomics and cheminformatics platforms.

AI/ML Focus

• DeepBiomarker and DeepBiomarker2 — deep-learning tools built on EHR data and social determinants of health to predict alcohol/substance-use-disorder and adverse-event risk in PTSD patients.
• T-DeepBiomarker, which integrates drug-target information into deep-learning models to predict adverse events and identify repurposing candidates in Alzheimer's disease with psychosis.

Representative Publications

• DeepBiomarker2 for predicting alcohol/substance-use-disorder risk in PTSD patients using EHR and social determinants of health. J. Pers. Med., 2024.
• Deep-learning models integrating drug-target information to predict adverse events in comorbid PTSD/alcohol-use disorder. Biomedicines, 2024.
• T-DeepBiomarker applied to identify potential medications for Alzheimer's disease with psychosis. Int J Mol Sci, 2025.
• 7 total AI/ML publications, 2020–2026, including in Drug and Alcohol Dependence and Pharmaceuticals.

Grants

• NIH/NIMH R01MH116046 (2018–2028), mPI — MODEL-AD+P; deep-learning models to identify medications for psychosis in Alzheimer's disease; developed DeepBiomarker.
• DOD W81XWH-15-2-0077 subcontract (2022–2024), Subcontract PI — Deep-learning/ML mining of EHR data to identify repurposing candidates for PTSD with substance-use disorder.
• NIH/NIDA P30 DA035778 (2014–2020), Co-I — Center of Excellence for Computational Drug Abuse Research; developed TargetHunter, HTDocking, and a BBB predictor.

Trainees

• Peihao Fan
• Oshin Miranda
• Chen Jiang
• Xiguang Qi
• Hongyi Zou
• Daniel Lee

Philip Empey, PharmD, PhD
 - Pharmacogenomics & Digital Health

Applies AI/ML to identify predictors of medication response using pharmacogenomics, and builds generative-AI chatbots and tools to support PGx-informed decisions and education.

AI/ML Focus

• AI/ML to identify predictors of medication response using pharmacogenomic (PGx) data.
• Generative-AI chatbots that help patients and providers apply PGx results in practice.
• AI tools to develop engaging, scalable online PGx education.

Representative Publications

Empowering personalized pharmacogenomics with generative AI solutions. PMID: 38447590 — Co-I; drove the PGx research questions and evaluation.

Trainees

• Olivia Romano (PhD student) — PGx prediction of delirium risk
• Grace Swatsworth (PharmD student) — AI coach for applying PGx results

Sandra L. Kane-Gill, PharmD, MS
 - Clinical Decision Support & Nephrotoxicology

Builds machine-learning clinical decision support systems to predict and manage drug-associated acute kidney injury, and evaluates AI-driven pharmacist interventions in pragmatic clinical trials.

AI/ML Focus

• Deep-learning/recurrent neural network models for dynamic, near-real-time acute kidney injury (AKI) risk prediction, with a focus on drug-associated AKI.
• Systematic evaluation of AI/ML models across pharmacy practice and nephrology, and NLP pipelines for mining clinical notes to inform decision support.

Representative Publications

• Systematic review and meta-analysis of 302 externally validated ML models for AKI risk classification across 95 studies. JASN, 2025.
• Recurrent-neural-network model for dynamic postoperative AKI risk and severity prediction, using integrated gradients for explainability. Surgery, 2023.
• Study protocol for MEnD-AKI, a multi-hospital cluster-randomized trial of a deep-learning CDSS paired with pharmacist telemedicine. Contemp Clin Trials, 2025.
• Commentary forecasting AI's impact on clinical pharmacy practice across the medication-use process. J Am Coll Clin Pharm, 2025.

Active Grants

• NIH/NIDDK R01DK121730 — Contact PI (MPI with Bihorac) — MEnD-AKI: multicenter implementation of a neural-network-based clinical decision support system for drug-associated AKI, paired with pharmacist-led telemedicine intervention.

Trainees

• Matthew Gray, PharmD, PhD
• McKenna K. Anderson, PharmD, MS
• Britney A. Stottlemyer, PharmD
• Nabihah Amatullah, PharmD
• Tiffany L. Tran, PharmD