Kedan He
Associate Professor
Physical Sciences
Eastern Connecticut State University
Biography
I am an Associate Professor of Chemistry at Eastern Connecticut State University, Connecticut’s only public liberal arts university and a primarily undergraduate institution within the Connecticut State University system. I earned my M.S. in Materials Science from the University of Alabama in Huntsville and my Ph.D. from the University of Georgia’s Center for Computational Quantum Chemistry. Trained as a computational chemist, I now focus on bridging computational chemistry and cheminformatics, applying machine learning and molecular modeling to study chemical information and drug discovery. After earning tenure and promotion in 2024, I have continued to expand undergraduate research opportunities at Eastern, mentoring students who have co-authored publications and presented their work at national conferences. I am an active member of the MERCURY consortium and a recipient of the 2023 SRP program grant, partnering with Oak Ridge National Laboratory’s leading computing facility. In the classroom, I strive to create inclusive and interdisciplinary learning experiences that connect chemistry to real-world contexts in the life, environmental, and health sciences. My work aims to advance data-driven discovery in chemical research while fostering the next generation of scientists through transformative undergraduate experiences.
Degrees Earned
Ph.D., Computational Chemistry, University of Georgia, Athens, GA (2009-2015) Thesis: "High-level Ab Initio Studies of Amino Acid Conformations, Rotamerizations, and Tunneling Dynamics" Advisor: Dr. Wesley D. Allen M.S., Materials Science, University of Alabama in Huntsville, AL (2007-2009) B.S., Polymer Materials and Engineering, Hainan University, China (2003-2007)
Research Areas
Chemistry; Machine Learning/AI
Research Interests
Cheminformatics and Drug Discovery: I develop computational workflows that integrate molecular descriptors, pharmacological fingerprints, and protein sequence embeddings to predict structure-activity relationships and identify bioactive compounds. Recent projects focus on multi-task learning approaches for CNS drug identification, Alzheimer's disease inhibitor discovery, and detecting emerging designer drugs and synthetic cathinones. Machine Learning in Chemical Sciences: I apply supervised and unsupervised ensemble learning methods to high-dimensional chemical data, including spectroscopy data and large-scale databases like DrugBank and PubChem. My work explores how AI can extract meaningful patterns from complex molecular datasets to accelerate therapeutic development. Computational Chemistry Applications: Leveraging my quantum chemistry background, I use DFT simulations and molecular modeling to generate training data for machine learning models, bridging traditional computational chemistry with modern AI techniques. Undergraduate Research and Education: A passionate advocate for undergraduate research in computational sciences, I mentor students through the complete research cycle—from literature review and data curation to algorithm development and scientific communication. I'm committed to making advanced computational research accessible in a liberal arts setting, helping students develop both domain expertise in chemistry and computational literacy. Curriculum Innovation: I design courses that connect chemistry to biological, environmental, and health science contexts, emphasizing data analysis, molecular modeling, and real-world applications. My teaching philosophy centers on fostering scientific curiosity while building practical computational skills that prepare students for interdisciplinary careers in modern chemical sciences.
Topical Areas
Artificial Intelligence and Intelligent Systems; Biochemistry and Molecular Biology; Computer Science; Health Sciences; Informatics, Analytics and Information Science; Other Biological Sciences; Other Computer and Information Sciences
Research Synergy
Artificial Intelligence and Intelligent Systems: At the core of my work is the application of machine learning and deep learning to chemical and biological problems. I develop multi-task neural networks, ensemble learning models, and leverage language model architectures to predict molecular properties and drug-target interactions. My projects explore how AI can extract meaningful patterns from high-dimensional molecular data—moving beyond traditional rule-based approaches to data-driven discovery that scales with the exponential growth of chemical databases. Computer Science: My research requires fluency in algorithm development, data structures, and computational optimization. I design and implement custom computational workflows that handle feature engineering, model training, and validation pipelines. This includes developing efficient methods for processing large-scale datasets, implementing parallel computing strategies, and optimizing code performance—skills that directly align with core computer science principles. My partnership with Oak Ridge National Laboratory's computing facility further extends this work into high-performance computing environments. Health Sciences: The translational potential of my research directly addresses health science challenges. By predicting pharmacological properties, identifying potential therapeutics for Alzheimer's disease and CNS disorders, and detecting emerging synthetic drugs, my work contributes to more efficient drug discovery pipelines and public health protection. Understanding how small molecules interact with biological targets has direct implications for disease treatment and patient outcomes—bridging the gap between computational predictions and clinical relevance. Informatics, Analytics, and Information Science: Cheminformatics sits squarely within the broader informatics landscape. My work involves curating, standardizing, and mining large-scale chemical and biological databases (DrugBank, PubChem), developing molecular fingerprints and descriptors, and creating information systems that enable efficient chemical data retrieval and analysis. I apply statistical modeling and data analytics to extract chemical knowledge from structured and unstructured data sources, addressing fundamental information science challenges around data quality, integration, and knowledge representation. The Interdisciplinary Advantage: These disciplinary intersections aren't merely theoretical—they define my research methodology. A single project might involve: (1) curating chemical data from multiple sources (informatics), (2) developing machine learning models to predict biological activity (AI/computer science), (3) validating predictions against known therapeutic targets (health sciences), and (4) interpreting results through the lens of molecular interactions (chemistry/biochemistry). This integrative approach mirrors the collaborative, boundary-crossing nature of modern scientific discovery. Working at a primarily undergraduate institution, I've found that this interdisciplinary positioning is particularly powerful for student development. It exposes undergraduates to the reality that meaningful scientific problems rarely fit neatly into single disciplines, preparing them for careers that increasingly demand computational fluency alongside domain expertise.
Motivation
At my small state liberal arts university, and within my small department that primarily supports other programs, I am the only faculty member working at the interdisciplinary intersection of chemistry, data science, and computer science. This unique position allows me to mentor a small but highly motivated cohort of students who possess strong foundations in both biochemistry and data science—students who are eager to apply computational tools to meaningful scientific questions. However, due to the limited size of our institution and the service-oriented nature of my department, opportunities for intra-institutional collaboration are scarce, and our research resources are modest. These constraints make it challenging to offer students the breadth of exposure and technical infrastructure often found at larger research institutions. The opportunity to collaborate with the NSF NAIRR initiative would be transformative for both my students and my research program. Access to national-scale computational resources and mentorship from leaders in AI-driven science would significantly enhance the scope and impact of our research. It would also allow my students—many of whom come from underrepresented backgrounds—to participate in cutting-edge scientific work that they might not otherwise encounter, helping them build professional networks and see themselves as contributors to the broader scientific community. Through this collaboration, I also hope to deepen my expertise in computational methodologies, integrate these tools more fully into my teaching, and develop sustainable, long-term research partnerships that will continue to benefit my students and institution beyond the duration of the program. My ultimate goal is to empower students to see themselves as capable scientists—equipped with the skills, confidence, and support to pursue ambitious academic and professional goals in STEM.
Supervising Students Plan
I supervise students through a structured mentorship approach that emphasizes regular communication, collaborative tools, and clearly defined expectations. Currently I mentor both student participants in Fall 2025 semester in the form of honors theses (Ashton Croteau) and independent research (Yesli Linares-Lopez). Our supervision framework includes at minimum weekly one-on-one meetings as well as group meetings to provide consistent guidance and foster peer learning/collaboration. All meetings are documented through live notes shared on OneDrive that capture key discussion points, action items, and deliverables, ensuring accountability and maintaining a clear record of progress. We collaborate using a mix of in-person sessions and asynchronous remote communication via Slack. We are also granted access to Clemson University's Palmetto supercomputing cluster. During the 10-week SRP program, I anticipate closely supervising students in intensive, hands-on coding sessions held in person to accelerate technical skill development. A key advantage of this program structure is the enhanced troubleshooting support it provides—students receive guidance not only from me but also from field specialists and NSF project leaders who bring deep domain expertise and diverse problem-solving approaches. This multi-layered mentorship exposes students to how professionals in different roles tackle complex technical challenges, broadening their understanding of research methodologies and best practices. Such interactions with experienced researchers and project leaders also provide invaluable networking opportunities and insights into career pathways, helping students make more informed decisions about graduate programs or industry positions aligned with their interests. I also establish clear milestones and timelines for each phase of the project, providing formative feedback on students' written project reports, and supporting students' professional development through required oral and/or poster presentations at campus or regional research conferences. The collaboration with project leaders from NAIRR aligns with my mentoring objectives to encourage student engagement with broader scholarly communities, and potential co-authorship opportunities on publications when appropriate. This comprehensive approach is designed to help students develop both the technical competencies and professional skills essential to their academic and career trajectories.
Student Merit
I have had the privilege of working closely with both students and can attest to their preparedness and aptitude for this research opportunity. Ashton Croteau completed General Chemistry with me during the 2023-2024 academic year, and Yesli Linares-Lopez took the same two-semester general chemistry course with me in 2024-2025. Both students demonstrated excellent academic skills and an eagerness to learn beyond the limitations of the course curriculum. Both are majoring in biochemistry with a minor in bioinformatics, which equips them with the interdisciplinary background essential for these research projects. Ashton Croteau has completed one year of research collaboration with me in the form of independent research, where he took the initiative in identifying potential research topics, conducting literature searches and reviews, and participating in meaningful discussions that advanced the progress of the project. During his initial onboarding to the project, he independently wrote Python code to curate the DrugBank datasets according to the research objectives, demonstrating strong coding skills and an exceptional independent work ethic. Additionally, Ashton participated in the University of Connecticut's Summer Research Internship Program in Biological and Biomedical Sciences in summer 2025, further expanding his research experience and exposure to rigorous scientific inquiry. He is currently writing the project proposal for his honors thesis requirement, focusing on the methodology aspect of the project. His ability to work independently, think critically about research design, and synthesize complex scientific literature, combined with his proven track record in competitive research programs, makes him well-prepared for the intensive nature of the SRP program. Yesli Linares-Lopez started working with me this semester. While reviewing key literature for this project, she is also completing online collaborative Data Chemistry OLCC course modules on cheminformatics to strengthen her computational chemistry foundations. She has demonstrated strong initiative and intellectual curiosity about the research problems, asking insightful questions and quickly grasping sophisticated concepts. Her proactive approach to building relevant skills and her enthusiasm for tackling challenging problems position her well for success in this program. Both students possess the technical foundation, work ethic, and genuine interest in computational chemistry research that will enable them to make meaningful contributions to this project while gaining valuable research experience.
Lightning Talk Title
Drug-Drug Interaction Prediction Using Node2Vec Embeddings and Ensemble Learning
Keywords
Polypharmacy; Drug-Drug Interactions (DDIs); Multi-Target Drug Design; Computational Pharmacology; AI/Machine Learning in Drug Discovery; Disease Module Networks; Precision Medicine