A. Pairwise similarity between strains based on phenotypes under different conditions. For source data '2022_DeChiara_Nature_Ecology_&_Evolution' strains with >50% missing values were filtered. Then the missing values were imputed by KNN. The similarity was based on the imputed data. For source data '2018_Peter_Nature' there are no missing values. So all strains are reserved. The similarity was assessed by Pearson correlation coefficient between strains. Hierarchical clustering was further applied at the strain level. The exact comparison pairs (the x-axis and y-axis values) and their corresponding coefficient value (the z-axis value) can be obtained by hovering the mouse on top of the heatmap