Fig. 2: Ploidy performance across models, inputs, and datasets.
From: A foundational model for in vitro fertilization trained on 18 million time-lapse images

Dot plots represent model performance based on AUROC for ploidy prediction on ES (Embryoscope) and ES+ (Embryoscope + ) microscopes. Performances are aggregated across 4 replicates (four-fold cross validation) (n = 4). Euploid vs. aneuploid classification: a image input; Weill ES, p = 0.84; 2020 Weill ES + , p = 2.1e-24; Spain ES + , p = 2.3e-24; Florida ES + , p = 0.69; 2021+ Weill ES + , p = 3.2e-21. b image + age input; Weill ES, p = 0.07; 2020 Weill ES + , p = 1.7e-16; Spain ES + , p = 2.9e-19; Florida ES + , p = 0.14; 2021+ Weill ES + , p = 4.5e-14. c video input; Weill ES, p = 2.6e-23; 2020 Weill ES + , p = 6.3e-25; Spain ES + , p = 4.1e-27; Florida ES + , p = 1.7e-19; 2021+ Weill ES + , p = 1.0e-22 d video + age input; Weill ES, p = 0.79; 2020 Weill ES + , p = 0.99; Spain ES + , p = 0.99; Florida ES + , p = 0.69; 2021+ Weill ES + , p = 0.99. For euploid vs. complex aneuploid classification: e image input; Weill ES, p = 2.7e-20; 2020 Weill ES + , p = 9.2e-25; Florida ES + , p = 4.3e-21; 2021+ Weill ES + , p = 4.8e-26. f image + age input; Weill ES, p = 8.6e-10; 2020 Weill ES + , p = 0.99; Florida ES + , p = 2.0e-11; 2021+ Weill ES + , p = 6.6e-12. g video input; Weill ES, p = 1.7e-17; 2020 Weill ES + , p = 3.4e-22; Florida ES + , p = 0.99; 2021+ Weill ES + , p = 1.2e-22. h video + age input; Weill ES, p = 0.07; 2020 Weill ES + , p = 0.73; Florida ES + , p = 0.58; 2021+ Weill ES + , p = 0.99. Asterisks represent statistical significance with p < 0.05, where statistical significance is determined by performing a one-way ANOVA test followed by a Tukey HSD test. For all subplots, error bars show mean values +/- SEM. Source data are provided as a Source Data file.