An Indiana University Bloomington researcher will co-lead a $4 million award from the National Institutes of Health’s National Cancer Institute to expand access to computational biology tools for conducting simulations in search of new cancer therapies, as well as modeling other biological processes.
Paul Macklin, an associate professor at the IU Luddy School of Informatics, Computing and Engineering in Bloomington, is a primary investigator on an award from the NCI’s Informatics Technology for Cancer Research program, which develops critical tools and resources to improve the acquisition, analysis, visualization and interpretation of data across the cancer research continuum.
With the support of IU’s high-performance computing resources, Macklin and collaborators will work to integrate two powerful research tools into a single cloud-based platform open to cancer researchers around the country and the world.
The first tool is PhysiCell, an “agent-based modeling” system developed by Macklin’s lab to simulate how cells grow, migrate and communicate in cancer and disease. Over the past several years, PhysiCell has been used to investigate breast cancer, colorectal cancer and pancreatic cancer, as well as cancer metabolism and immunotherapy.
The other tool, CoGAPS, is led by grant co-leader Elana Fertig, incoming director of the Institute for Genome Sciences at the University of Maryland. CoGAPS analyzes large-scale genomics data — including “spatial transcriptomics,” or gene sequencing that captures cells’ physical positions — to discover cell identities, behaviors and interactions in tissue samples. The software then uses these data to generate rules for use in simulations in PhysiCell.
Integrating these systems will produce a tool that can quickly create agent-based modeling simulations based on spatial transcriptomics data. Researchers can apply these simulations to understand how changes to biological systems — like cancer stage or therapies — impact cancer cells.
“PhysiCell enables intuitive agent-based modeling: Scientists ‘encode’ their knowledge of cell biology as cell rules that stipulate how cellular signals drive changes in cell behaviors, such as cell cycling, death, migration, secretion or phagocytosis,” said Macklin, who is also a researcher with the IU Simon Comprehensive Cancer Center.
“However, it’s currently difficult to quantitatively connect these rules to single cell and spatial transcriptomics data. This project will allow agent-based modeling to reach its potential.”
For example, a researcher might use the system to quickly build a virtual model of a real patient’s pancreatic cancer, then simulate thousands of potential drug compounds to predict which may have a therapeutic effect.
The power of computational biology is the ability to run thousands of virtual experiments on simulated patients to grasp a compound’s potential efficacy and safety with no risk to a patient, Macklin said.
The project is also a major milestone in the development of PhysiCell. After years advancing the system to solve specific challenges in cancer research, IU researchers can now focus on refining the tool’s usability and flexibility in support of the national cancer research data analysis infrastructure.
Macklin said PhysiCell has already grown significantly in usability, noting that many improvements stem from the work of undergraduate and graduate students in the IU Math Cancer Lab or in their coursework.
For example, a recent update to the system introduced a “novel cell behavior grammar”: a simplified programming language to help scientists build simulated biological systems without traditional code.
PhysiCell currently has hundreds of users across the globe. It’s not uncommon to learn about others using their tools through citations in research papers, Macklin said.
“Most of the time, we never know people are using PhysiCell until we see the citation,” he said. “It’s very gratifying since it means we’ve created something useful enough that they’re not calling us up for tech support.”
The new grant will support not only further development of the tools but also continued outreach and training, such as virtual hack-a-thons and YouTube tutorials.
In addition to improved performance, Macklin said PhysiCell’s use has grown among scientists due to its strong focus on reducing the “pain points” in computational biology.
“Computer science tends to put too much emphasis on the performance side,” he said. “However, if you look at the whole timeline required to do computational biology — not just the simulation speed but also the time to train on a system and convert experiments into code — then you’re not really saving a lot of time. Speeding up simulations focuses on the smallest sliver of the problem.”
Conversely, if the tools are easier to use — with low-code environments, graphical user interfaces and strong documentation — then “you’re truly starting to slash the amount of time required to solve big problems in biology,” he said.
Robert Quick of IU’s Pervasive Technology Institute is also a co-investigator on the grant. The team works with a broad network of collaborators at IU, Johns Hopkins University, Oregon Health and Science University and the University of Maryland at Baltimore.
