Author response image 13. Resampling analysis.
CD8+ TCell populations were sampled 100 times for different sampling sizes (10 to 1000 with a step of 10). For every simulation, we reclassified the patient into “Active”/”Transition”/”Exhausted” using the approach described in the methods. This analysis shows the robustness of the patient classification methodology with most of the patients requiring a very small sample size in order to obtain at least 95% of consistent classifications.