Bridging the Gap: How Research Computing Powers the Future of Molecular Informatics

At the University of Rhode Island, the move toward data-intensive research is led by scientists who maintain a dual fluency in biological systems and the high-performance computing environments used to analyze them. For Dr. Christopher Hemme, Director of the Rhode Island INBRE Molecular Informatics Core, URI’s research computing infrastructure is the backbone of a mission to democratize data science for biomedical researchers across the region.

A Multidisciplinary Hub for Discovery
Dr. Hemme leads the data science core within the Rhode Island IDeA Network of Biomedical Research Excellence (INBRE), a program designed to complement instrumentation facilities with advanced computational support. His work is remarkably diverse, spanning bioinformatics, data management, and even 3D science visualization.
The Core’s impact is felt across a wide array of specialized projects. Currently, Dr. Hemme is working with Dr. Jamie Ross and Dr. Giuseppe Cappotelli to develop a pipeline for analyzing mitochondrial damage in mouse models related to aging and Alzheimer’s disease. Simultaneously, he is collaborating with Dr. Nisa Ghonem on proteomics and metabolomics related to liver disease, and utilizing the NIH’s “All of Us” database to conduct large-scale cohort studies and genome-wide association analyses (GWAS).
Beyond the bench, the Core is also pushing the boundaries of training and education. “We partner with IT services to build VR apps for training and research,” Dr. Hemme explains. This includes the HERBAL project with Dr. David Rowley that has created a virtual medicinal garden, coral reef, and research laboratories to train high school students in the methods of natural product chemistry.

Unity: From Unreliable Servers to High-Performance Power
The landscape of research computing at URI has shifted dramatically during Dr. Hemme’s tenure. He recalls a time when the network relied on small, unreliable clusters and localized servers that weren’t powerful enough for modern genomic demands. The development of the Unity cluster over the last decade has been a game-changer.
“Having Research Computing built up has been very helpful, especially as we get into AI,” says Dr. Hemme. “Unity, in particular, is a great resource. If you’re not familiar with the command line, the interactive resources available through Unity make it much more accessible to researchers than it would have been in the past.”
For the INBRE network, which includes researchers at smaller undergraduate institutions (PUIs) with limited local resources, the ability to gain affiliate status on Unity is critical. It allows scientists across Rhode Island to access the same high-level computing power as those at major research universities, simplifying workflows and removing the burden of server maintenance.

Navigating the Multi-Omics Era
As sequencing instruments and mass spectrometers generate increasingly massive datasets—often reaching the terabyte range—Dr. Hemme sees URI’s Open Storage Network (OSN) as fundamental for archiving and data management.
The complexity of the work is also evolving. Dr. Hemme is currently preparing students for a “new era” of research that moves beyond bulk sequencing into spatially resolved multi-omics. This requires integrating noisy data from different molecular levels—transcriptomes, proteomes, and metabolomes—and correlating them with health outcomes.
“It’s a nightmare to integrate all of it,” Dr. Hemme admits. “It requires significant integration strategies to get these different noisy omics levels integrated together. I tell my students: everything is regression. If you learn regression, you can do a lot of data science.”

The AI Frontier in Pharmacology
One of the most exciting growth areas for the Core is the application of Artificial Intelligence. Dr. Hemme is currently working on an “AI agent” for pharmacology, built to access generalist databases and repositories with less structured data than traditional sources like PubMed. This agent aims to automate workflows and feed a broad array of data into scientific analysis.
He is also tackling the challenges of AI in structural biology, working to build a training set that can recognize components in complex multi-figure illustrations. “It’s a challenge because there’s so much variety in images—from old pencil drawings to modern, stereoscopic illustrations,” Dr. Hemme notes. “We’ve pulled about 150 images so far to help the model find what’s specifically interesting about a structure.”

Mentorship and the “Black Box” Problem
Despite the high-tech tools, Dr. Hemme insists that researchers remain deeply involved in the process. He works to ensure that bioinformatics tools do not become “black boxes” where data goes in and results come out without critical oversight.
“Even if someone comes to me and lets me do their analysis, I don’t necessarily know their biological system,” he says. “The researcher needs to be able to determine if the pipeline is working correctly and if the results make sense. The more familiar you are with these resources, the better we can communicate.”

Advice to the URI Community
Dr. Hemme’s advice for faculty and students is to move past the intimidation factor and embrace the scale of what is possible at URI.
“I’d strongly encourage anyone to use these resources. In a lot of cases, people don’t know what’s possible, so they try to do things on laptops that don’t have the resources,” he notes. “It’s a good investment to learn a little bit of what’s happening under the hood. It makes your workflows more efficient and your science more robust.”

Written by Leann Biancani and Cecile Cres.