A data-centric computational research group at the University at Buffalo leveraging heterogeneous biomedical datasets to unravel drug mechanisms, predict adverse effects, and accelerate therapeutic discovery.
The Falls Informatics Group integrates computational chemistry, machine learning, and clinical data science to bridge the gap between molecular mechanisms and patient outcomes. Based in the Department of Biomedical Informatics at the Jacobs School of Medicine and Biomedical Sciences, our work spans from atom-level drug–protein interactions to population-level electronic health record analyses.
We develop and apply novel algorithms and platforms—most notably the CANDO multiscale drug discovery platform—to repurpose existing therapeutics, predict polypharmacy risks, and identify new treatment candidates for infectious diseases, cancers, and substance use disorders.
Meet the TeamThe CANDO platform integrates chemical, structural, and pharmacological data to predict drug–proteome interactions across the entire human proteome.
Mining electronic health records and Medicaid claims data to identify disease risk factors, prescription trends, and adverse drug event signals.
Deep learning and graph neural networks for de novo drug design, binding affinity prediction, and RNA editing identification.
Applications in non-small cell lung cancer, COVID-19, opioid use disorder, and influenza with wet-lab validation collaborators.
Our research program is organized around four interconnected thrusts that span from computational algorithm development to clinical translation.
Development and enhancement of the CANDO platform for multiscale drug–proteome interaction modeling and therapeutic repurposing using machine learning and structural bioinformatics.
Predicting and mitigating adverse drug reactions arising from polypharmacy using integrative computational approaches across molecular and clinical scales.
Electronic health record and claims data analytics to characterize prescribing patterns, disease risk, and treatment outcomes in real-world patient populations.
Rapid therapeutic candidate identification for emerging pathogens including SARS-CoV-2, influenza, and other infectious agents via computational screening.
Multiscale analysis of drug combination efficacy targeting oncogenic driver mutations, with a focus on KRAS-mutant non-small cell lung cancer.
Population-level informatics approaches to understand opioid, benzodiazepine, and alcohol co-prescribing patterns and their clinical consequences.
Research supported by NIH, NLM, NIDA, and private foundations.
The University at Buffalo received a renewal of its NIH/NCATS Clinical and Translational Science Award, totaling $29.2M over six years. Dr. Falls contributes biomedical informatics expertise to the award's research infrastructure and training cores.
Dr. Falls was appointed as Affiliated Faculty in the University at Buffalo Institute for Artificial Intelligence and Data Science, reflecting the group's growing engagement with machine learning for biomedical discovery.
Dr. Falls was appointed as Adjunct Research Assistant Professor at Roswell Park Comprehensive Cancer Center, formalizing a collaborative relationship supporting oncology informatics and computational drug discovery.