[{"id":3648,"date":"2026-10-09T15:31:08","date_gmt":"2026-10-09T15:31:08","guid":{"rendered":"https:\/\/blogs.vcu.edu\/cctr\/?p=3648"},"modified":"2026-10-09T15:31:10","modified_gmt":"2026-10-09T15:31:10","slug":"how-a-small-nih-pilot-at-vcu-helped-launch-a-precision-medicine-company","status":"publish","type":"post","link":"https:\/\/blogs.vcu.edu\/cctr\/2026\/10\/09\/how-a-small-nih-pilot-at-vcu-helped-launch-a-precision-medicine-company\/","title":{"rendered":"How a small NIH pilot at VCU helped launch a precision medicine company"},"content":{"rendered":"\n

A seed grant through the Wright Center offered crucial support for TACIT \u2014 an AI tool for reading human tissue cell by cell \u2014 and catalyzed a spinout that has since expanded to build virtual tissues for precision medicine.<\/em><\/p>\n\n\n\n

The human body is not flat. It is a crowded, three-dimensional world, built from trillions of cells that live side by side, lean on their neighbors and constantly trade signals. A cancer cell is dangerous not only because of what it is, but because of the company it keeps and where it sits. To understand disease \u2014 and to beat it \u2014 scientists need to understand not only what each cell is, but where it lives, who it communicates with, and how those relationships change during health and disease.<\/p>\n\n\n\n

For a long time, they could not. Now, thanks in large part to a small pilot grant at Virginia Commonwealth University, they can.<\/p>\n\n\n\n

One pivotal chapter began in 2024, when Jinze Liu, Ph.D., won a <\/a>Translational Science Pilot Grant<\/a> from the <\/a>Wright Regional Center for Clinical and Translational Science<\/a>, the home of Virginia’s Clinical and Translational Science Award. Liu is a professor in the Department of Cellular, Molecular and Genetic Medicine<\/a> at the VCU School of Medicine<\/a>, co-lead for Health Informatics at the Wright Center, and the director of Bioinformatics Shared Resource<\/a> at Massey Comprehensive Cancer Center at VCU. Her <\/a>funded project<\/a> took aim at one of the hardest problems in a fast-growing field called spatial biology.<\/p>\n\n\n\n

Spatial biology lets scientists <\/a>map the location and interactions of cells inside tissue<\/a>, keeping the context that older methods threw away. But there was a catch. Sorting millions of cells into their correct types was <\/a>painfully slow, often done by hand by expert pathologists<\/a>. Identifying cells in a single tissue sample could take more than a month.<\/p>\n\n\n\n

That is the wall Liu and her pilot-grant team set out to break. Joining her as co-investigators were Katarzyna Tyc, Ph.D., an assistant professor in the Department of Biostatistics in VCU\u2019s School of Public Health<\/a>, who developed the algorithm along with Khoa Huynh, then a graduate student in biostatistics, and Kevin Matthew Byrd, D.D.S., Ph.D., an <\/a>assistant professor of oral and craniofacial molecular biology at the VCU School of Dentistry<\/a> and a fellow Massey member who supplied tissue data and expert cell labeling. The pilot also supported the doctoral training of Huynh, who implemented the TACIT algorithm<\/a> and earned his Ph.D. in biostatistics in 2025. Together they published TACIT \u2014 short for Threshold-based Assignment of Cell Types from Multiplexed Imaging Data. While TACIT solved the immediate challenge of rapidly annotating spatial biology data, it also established the computational foundation for a much broader platform capable of reconstructing virtual tissues and modeling disease at cellular resolution.<\/p>\n\n\n\n

TACIT reads high-resolution images of tissue and <\/a>sorts cells by type in minutes instead of months<\/a>, leveraging machine learning algorithms. To train and test it, the team used data from 5 million cells and 51 cell types across three very different parts of the body: the brain, the intestine and glandular tissues<\/a>. In tests against three widely used methods, TACIT <\/a>lined up more closely with expert annotations, scaled to much larger datasets, and was especially good at spotting rare cell types that other tools often miss<\/a>.<\/p>\n\n\n\n

“We’re using artificial intelligence to increase efficiency and also the accuracy of diagnosis,” <\/a>Liu said<\/a>.<\/p>\n\n\n\n

\n