Fig. 5: Burden heritability and genetic correlation.

A The distribution of burden heritability point estimates for ultra-rare (minor allele frequency <1 × 10⁻⁴) pLoF variants and damaging missense variants (i.e., the int1 damaging missense variants) across 591 multi-organ imaging traits (including 36 abdominal MRI traits, 82 CMR traits, 101 Regional Brain Volumes, 110 DTI traits, 82 resting-state ICA-based fMRI traits, 90 parcellation-based fMRI traits, and 90 parcellation-based fMRI traits) that had a sample size larger than 10,000 (5 abdominal imaging traits were excluded due to insufficient sample size). Burden heritability was estimated based on the summary statistics of each variant-level tests and all the individuals (N = 54,365) were covered. In the boxplots shown, each box represents the distribution of burden heritability estimates across a certain imaging phenotype category for pLoF or damaging missense variant groups. The central line within each box denotes the median value; the upper and lower bounds of the box correspond to the third (Q3) and first (Q1) quartiles, respectively. Whiskers extend to the most extreme data points within 1.5 times the interquartile range (IQR) from the box, and points beyond the whiskers indicate outliers. The dashed horizontal line marks the null expectation. B Burden genetic correlations of ultra-rare (minor allele frequency <1 × 10⁻⁴) pLoF variants for selected imaging phenotypes and other complex traits. P-values are derived from two-sided tests based on z-scores. Only top hits are presented here. Specifically, we first selected seven trait pairs from pairs that had ten smallest P-values among all the trait pairs and additionally included three pairs that had smallest or second smallest P-values for regional brain volumes and abdominal MRI traits, which resulted in nine imaging traits across all MRI categories and seven Genebass traits. Filled circles indicate the selected top hits (P < 0.0032) while filled triangles indicate other pairs that passed a nominal significance threshold at P < 0.05. “Phenotype_ID” was used to define the selected imaging traits. Specifically, abdominal fat ratio from abdominal MRI traits, left atrium maximum volume (LAV_max), regional radial strain (Err_AHA_13) and peak circumferential strain (Err_AHA_4) traits from CMR, left and right insular volumes, functional activity trait from resting fMRI (ICA-based, Net100_Node55) traits, Visual2-Auditory network functional activity trait from task fMRI (parcellation-based), and Somatomotor-Dorsal-Attention network functional activity trait from resting fMRI (parcellation-based) traits were shown here (bottom-to-top order). Supplementary Data S1 includes more details of these phenotypes.