Related Experiment Video
Updated: Jun 5, 2026
Predictive Immune Modeling of Solid Tumors
Published on: February 25, 2020
A community machine learning challenge to predict the effects of gene perturbations on T cell differentiation for
Jiaqi Zhang1,2, Marc A Schwartz3, Mohammed Mutaher3
1Eric and Wendy Schmidt Center, Broad Institute of MIT and Harvard, Cambridge, 02142, MA, USA.
Abstract:
Perturbations of genes with functional importance in T cells could be used to change the distribution of CD8 T cell states to enhance anti-tumor functions for cancer immunotherapies. We launched a world-wide computational challenge to predict the effects of gene perturbations and to devise objective functions for prioritizing gene perturbations that lead to desired T-cell state distributions. We supported the challenge by generating a single-cell Perturb-seq dataset profiling the effect of knocking out 73 individual expert-defined genes in T cells transferred into a mouse melanoma model. We compared the top algorithms developed by participants, and found that performance was primarily determined by the prior data used for gene feature representation, with perturbational data derived features, proving most effective. Experimental validation of the top 61 genes nominated by the algorithms revealed that perturbation of Ndufv2 and Dimt1 reached the defined objective and biased T cell differentiation toward desired states.
Insights
Gene perturbations in T cells can enhance anti-tumor immunity for cancer immunotherapies. A computational challenge identified Ndufv2 and Dimt1 gene perturbations that effectively bias T cell differentiation toward desired anti-cancer states.
Area of Science:
- Immunology
- Computational Biology
- Genetics
Background:
- T cell state manipulation is crucial for advancing cancer immunotherapies.
- Targeting specific genes can alter T cell populations to improve anti-tumor responses.
Purpose of the Study:
- To develop computational methods for predicting gene perturbation effects on T cell states.
- To identify objective functions for prioritizing gene perturbations that enhance anti-tumor functions.
Main Methods:
- Generated a single-cell Perturb-seq dataset for 73 gene knockouts in T cells within a mouse melanoma model.
- Launched a global computational challenge to predict gene perturbation outcomes.
- Compared participant algorithms, focusing on feature representation from prior data.
Main Results:
- Perturbational data-derived features were most effective for algorithm performance.
- Experimental validation confirmed that Ndufv2 and Dimt1 perturbations achieved desired T cell state distributions.
- Top algorithms successfully nominated genes for targeted T cell differentiation.
Conclusions:
- Computational approaches, particularly those using perturbational data features, can effectively predict gene perturbation impacts on T cells.
- Ndufv2 and Dimt1 represent promising targets for engineering T cells to enhance anti-tumor immunity in cancer therapy.
