Fig. 5: RNAPath accuracy, measured by computing the Pearson correlation coefficient (r) between expression prediction and bulk RNA-seq, across 23 GTEx tissues. | Nature Communications

Fig. 5: RNAPath accuracy, measured by computing the Pearson correlation coefficient (r) between expression prediction and bulk RNA-seq, across 23 GTEx tissues.

From: Self-supervised learning for characterising histomorphological diversity and spatial RNA expression prediction across 23 human tissue types

Fig. 5

The number of test samples used to derive statistics is reported next to each model on the x-axis. The violin plots depict the distribution of Pearson correlation coefficients, with the inner box showing their median and interquartile range (25th and 75th percentile) and the tails ranging in the interval (Q1 − 1.5 × IQR, Q3 + 1.5 × IQR), across the tested genes (cross-tissue average = 11,327).

Back to article page