Extended Data Fig. 2: Comparison between STORIES and its linear counterpart.
From: STORIES: learning cell fate landscapes from spatial transcriptomics using optimal transport
(A) In the boxplots, the center line, box limits, and whiskers denote the median, upper and lower quartiles, and 1.5× interquartile range, respectively. All scores are reported for n = 10 initialization seeds. Gene expression prediction score (top) and spatial coherence score (bottom) in early test sets (orange) and late test sets (green) across the three datasets. Scores are reported across seven values of quadratic weight α parameter, including the linear method (α = 0, light orange/green); (B) Wasserstein distance score in early test sets (orange) and late test sets (green) across the three datasets. Scores are reported across seven values of quadratic weight parameter α, including the linear method (α = 0, light orange/green); (C-E)Visual representation of the optimal transport matching involved in the loss of the linear method (top) and STORIES (bottom), for the three benchmark datasets. In each dataset, the left slice displays two cell types and the right slice displays the cells they are matched with at the following time point. The shown examples are: (C) IMN and rIPC2 cells in the axolotl dataset; (D) optic vesicle and Polster cells in the zebrafish dataset; (E) lung and liver cells in the mouse dataset.
