Fig. 3: Benchmarking zero-shot performance across multiple datasets. | Nature Communications

Fig. 3: Benchmarking zero-shot performance across multiple datasets.

From: PhenoProfiler: advancing phenotypic learning for image-based drug discovery

Fig. 3

a Performance evaluation of different models using leave-plates-out validation. b Performance evaluation of different models using leave-dataset-out validation. c Performance evaluation of different models using out-of-distribution validation. The test plates include BR00115125-BR00115134 (cpg0001) and SQ00014812-SQ00014816 (cpg0004). FoE: Folds of Enrichment; MAP: Mean Average Precision; BBBC036: CDRP-BIO-BBBC036; BBBC037: TAORF-BBBC037. Source data are provided as a Source Data file.

Back to article page