Fig. 5: Quantitative analysis of the phenotype correction strategy in PhenoProfiler.
From: PhenoProfiler: advancing phenotypic learning for image-based drug discovery
a The conceptual motivation for the design of PCs. Phenotypic differences between treated- and controlled- wells capture the treatment response. b Ablation experiments demonstrating the impact of PCs across three datasets, showing a consistent increase in the FoE with the inclusion of PCs. c Sensitivity analysis of the hyperparameter α in the PCs. d UMAP visualizations of feature representations generated by PhenoProfiler, with and without PCs, evaluating the harmonization of well-level features using IMAD. PCs: Phenotype correction strategy; IMAD: Inverse Median Absolute Deviation; FoE: Folds of Enrichment; MAP: Mean Average Precision; UMAP: Uniform Manifold Approximation and Projection. Source data are provided as a Source Data file. Created in BioRender. Song, Q. (2025) https://BioRender.com/y4tl93e.
