New research develops a computational method that allows to interpret T cell states, in cancer and infection, from single-cell transcriptomics data in the context of stable reference atlases.
Single-cell RNA sequencing technologies are revolutionising biomedical research by enabling exploration of disease-associated cell states. However, no consensus exists on how to define cell states. Lack of consistent cell state definitions limits scientists’ ability to “connect the dots” across studies and diseases.
Research led by Santiago Carmona and first-author Massimo Andreatta has developed a computational framework, ProjecTILs, which addresses this issue. A computational biology tool, ProjecTILs enables interpretation of T cell states from scRNA-seq data using reference atlases as stable, curated systems of coordinates and can help researchers interpret their scRNA-seq experiments. One notable advantage of ProjecTILs over currently available methods is that the reference atlas remains unaltered upon projection of new datasets. Beyond the interpretation of new data, in terms of known annotated cell subtypes, it can also aid the discovery of novel states that deviate from the reference e.g. as a consequence of a genetic alteration or tissue adaptation.
In a meta-analysis of over 100 patient and mouse tumors, findings published in Nature Communications also revealed a significant conservation of tumor-infiltrating T cell states across cancer types and species.
This research was supported through Santiago Carmona’s Ambizione Fellowship from the Swiss National Science Foundation. Contributing scientists of this project are Jesus Corria-Osorio (UNIL), George Coukos (Head of the Department of oncology and Director of the Ludwig Lausanne branch) and Rafael Cubas and Soren Muller (Genentech).