Extended Data Fig. 1: Details about MENTR ML.

(a) Comparisons between required datasets for MENTR ML models and conventional eQTL study. In MENTR ML models, CAGE transcriptome data is only required. Existing large-scale CAGE transcriptome, such as FANTOM5 datasets, can be used. In eQTL study, transcriptome data for the tissue and genotypes from the same individuals are required for estimating mutation effects for each transcript (βeQTL) in a tissue. (b) Workflow of MENTR ML training and evaluation. See the details in the Methods section.