Fig. 1: Characteristics of samples in the pilot phase of PigGTEx project.

a, Clustering of 7,095 RNA-seq samples based on the normalized expression (log₁₀-transformed TPM) of 6,500 highly variable genes, defined as the top 20% of genes with the largest s.d. of TPM across samples. b, The same sample clustering as a but based on normalized alternative splicing values (PSI) of 6,500 highly variable spliced introns, defined as the top 13% of spliced introns with the largest s.d. of PSI across samples. c, Principal component analysis of samples based on 12,207 LD-independent (r² < 0.2) SNPs. The left panel is for whole-genome sequencing samples (n = 1,602) in the PGRP, while the right one is for RNA-seq samples (n = 7,008) with successful genotype imputations. d, Sample sizes of 34 tissues, cell types and organ systems (all referred to as ‘tissues’) used for molQTLs mapping. e, Clustering of 34 tissues based on the median expression of all 31,871 Ensembl annotated genes (v100) across samples within tissues, representing embryo, endodermal, mesodermal and ectodermal lineages.