Fig. 2: SVEN improves tissue-specific gene expression prediction.

a SVEN outperforms state-of-the-art tools in predicting tissue-specific gene expression on held-out sequences that were excluded from the training dataset for testing. The predicted log10RPKM values were compared with the observed log10RPKM values from RNA-seq data for overlapped 218 tissues and cell lines with Enformer and ExPecto. Spearman correlations between the predicted and observed values were calculated for comparison. b The values predicted by SVEN versus the observed values for all the testing sequences in the 6 example tissues reported by ExPecto and Enformer. Source data are provided as a Source Data file.