IMMUNOLOGY2026™ Conference Recordings For Attendees-immgenT-GP: understanding T cells through gene programs

immgenT-GP: understanding T cells through gene programs

Video Transcription

Video Summary

The talk presented ImGene-TGP, a method for understanding T cells through gene programs rather than discrete clusters. Using semi-nonnegative matrix factorization, the team learned 200 sparse programs and validated them with nonlinear reduction. Many programs were lineage-specific, revealing canonical markers for CD8, Treg, and other T-cell states, while activation involved multiple programs rather than one universal switch. They also annotated programs linked to cytotoxicity, tissue residency, exhaustion, and helper T-cell states. Finally, they showed some programs could be predicted using surface proteins, suggesting potential for flow-cytometry-based validation.

Meta Tag

Date April 16, 2026 4:30 PM - 4:45 PM

Room 151

Session Resource Highlight: immgenT: A Community Reference Atlas for Mouse T Cells

Speaker Ziang Zhang

Track Computational and Systems Immunology (COMP)

Year 2026

Keywords

ImGene-TGP

T cell gene programs

semi-nonnegative matrix factorization

T-cell states

surface protein prediction

April 16, 2026 4:30 PM - 4:45 PM

151

Resource Highlight: immgenT: A Community Reference Atlas for Mouse T Cells

Ziang Zhang

Computational and Systems Immunology (COMP)

2026