Research Capacity Core – Integrated Informatics Services
Research Areas
- AI and Machine Learning for Biomedical Data Analysis
- Machine Learning Models for Precision Medicine
- AI and Machine Learning Training for the Biomedical Workforce
Scientific Achievements
- Precision medicine breakthroughs – Built predictive models to personalize drug response and define lab test reference intervals, improving diagnostic accuracy.
- AI-driven disease prediction tools – Applied machine learning to forecast outcomes in cancer, cardiovascular disease, epilepsy, and COVID-19, supporting earlier interventions.
- Advanced AI/ML methods for health data – Designed innovative algorithms for analyzing genomic, proteomic, pharmacogenomic, and clinical datasets to uncover biomedical insights.
- Training and workforce development – Led AI/ML courses and mentorship initiatives, preparing students and investigators to apply data science in health research.
Funding
RCMI Funding:
- NIH/NIMHD U54MD007600
Other funding obtained with RCMI support:
- 1OT2OD032581-02-235, NIH/AIM-AHEAD, Machine Learning Methods for Identifying Reference Intervals of Cardiometabolic Related Laboratory Tests for Hispanic Populations (MLM-RIHisp) | 2023–2025
- 1OT2OD032581-01-PP80, NIH/AIM-AHEAD, Identify Reference Intervals of Cardiometabolic Related Laboratory Tests Using Machine Learning (MLM) | 2022–2023
- 3U54MD007600-35S-35S2, NIH/NIMHD, Artificial Intelligence and Machine Learning to Health Disparities Research (AIML+HDR), Version 1,2 | 2021–2023
- 5R25CA240120-04, NIH/NCI, Artificial Intelligence and Machine Learning in Cancer Prevention and Control (AI/ML-CAPAC) Research | 2021–2022
Scientific Advance
Discovery of ancestry-specific variants associated with clopidogrel response among Caribbean Hispanics
Published in NPJ Genomic Medicine, Volume 10, 2025, PMCID: PMC11889249.
Published in NPJ Genomic Medicine, Volume 10, 2025, PMCID: PMC11889249.
The study analyzed DNA from 511 Puerto Rican cardiovascular patients taking clopidogrel (a blood thinner) to learn why some patients don’t respond well to the drug. A genome-wide scan found genetic variants tied to high “on-treatment platelet reactivity” (meaning the drug is less effective). Signals in the CYP2C19 region appeared mainly in people with more European ancestry, while several other variants (including OSBPL10 rs1376606, DERL3 rs5030613, RGS6 rs9323567, and one in UNC5C) were also implicated. The findings highlight that drug-response genetics can differ by ancestry and that commonly used markers may not predict clopidogrel response as well in patients with higher non-European ancestry.
U54 GM133807 (NIGMS); U54 MD007600 (NIMHD/RCMI); 1U54 MD010723 (NIMHD); R25 GM061838 (NIGMS).
