Fig. 2: Percentage of disease associations from a curated set of 253 known disease pairs that are assigned to the ten nearest neighbours based on cosine similarity for each disease embedding.
A data table with the exact values underlying the figure is given in Supplementary Table 9.
