Additional file 1 of Benchmarking foundation cell models for post-perturbation RNA-seq prediction

Supplementary Material 1: Supplementary Figure 1 – Subgroup analysis for Norman dataset Evaluation on the Norman dataset: Pearson delta metrics (y axis) for scGPT, scFoundation, Train Mean and Random Forest Regression with GO features (x axis, top) and for Random Forest Regression model with different features (x axis, bottom). Different subgroups (combo_seen0: none of the perturbations was present in the train set, combo_seen1: one of the perturbations was present in the train set, combo_seen2: both of the perturbations, but the combination was not seen in the train set, unseen_single: single perturbation, not seen in train set) are colour coded.