Quantifying concordant genetic effects of de novo mutations on multiple disorders

Hanmin Guo, Lin Hou, Yu Shi, Sheng Chih Jin, Xue Zeng, Boyang Li, Richard P. Lifton, Martina Brueckner, Hongyu Zhao, Qiongshi Lu

Abstract: Exome sequencing on tens of thousands of parent-proband trios has identified numerous deleterious de novo mutations (DNMs) and implicated risk genes for many disorders. Recent studies have suggested shared genes and pathways are enriched for DNMs across multiple disorders. However, existing analytic strategies only focus on genes that reach statistical significance for multiple disorders and require large trio samples in each study. As a result, these methods are not able to characterize the full landscape of genetic sharing due to polygenicity and incomplete penetrance. In this work, we introduce EncoreDNM, a novel statistical framework to quantify shared genetic effects between two disorders characterized by concordant enrichment of DNMs in the exome. EncoreDNM makes use of exome-wide, summary-level DNM data, including genes that do not reach statistical significance in single-disorder analysis, to evaluate the overall and annotation-partitioned genetic sharing between two disorders. Applying EncoreDNM to DNM data of nine disorders, we identified abundant pairwise enrichment correlations, especially in genes intolerant to pathogenic mutations and genes highly expressed in fetal tissues. These results suggest that EncoreDNM improves current analytic approaches and may have broad applications in DNM studies.

Introduction

Exome sequencing of parent-proband trios has revolutionized our understanding of the genetic architecture of developmental and psychiatric disorders. These studies have identified numerous deleterious de novo mutations (DNMs) that significantly contribute to disease risk. While many disorders appear clinically distinct, emerging evidence suggests they may share common genetic pathways. However, current analytical methods often struggle to capture the full extent of this overlap. Most existing strategies focus exclusively on genes that reach strict statistical significance within individual disorders. Given the high polygenicity and incomplete penetrance characteristic of these conditions, such "top-down" approaches likely overlook a vast landscape of shared genetic risk hidden within sub-threshold signals.

The EncoreDNM Framework

To address these limitations, we developed EncoreDNM (Enrichment correlation estimator for De Novo Mutations), a novel statistical framework designed to quantify shared genetic effects between two disorders. Unlike previous methods, EncoreDNM utilizes exome-wide, summary-level DNM data, incorporating information from all genes regardless of their individual significance level.

The EncoreDNM framework integrates gene mutabilities and mutation counts across variant classes to estimate enrichment correlation.

The core of EncoreDNM is a generalized linear mixed-effects Poisson regression model. The framework first establishes a baseline by calculating expected mutation rates for every gene based on sequence context. It then evaluates the deviation of observed mutation counts from these expectations. By treating these deviations as random effects, EncoreDNM calculates an enrichment correlation coefficient ($\rho$) to determine if genes over-mutated in one disorder show a similar trend in another. A critical feature of the model is the inclusion of an "elevation component," which accounts for technical noise and variability between different sequencing studies, ensuring more robust and calibrated results.

Data and Implementation

The performance of EncoreDNM was evaluated using a massive collection of exome sequencing data from over 45,000 parent-proband trios. The data covered nine distinct disorders:

  • Developmental Disorder (DD)
  • Autism Spectrum Disorder (ASD)
  • Schizophrenia (SCZ)
  • Congenital Heart Disease (CHD)
  • Intellectual Disability (ID)
  • Tourette Disorder (TD)
  • Epileptic Encephalopathies (EP)
  • Cerebral Palsy (CP)
  • Congenital Hydrocephalus (CH)

A control group of 1,789 healthy siblings of ASD patients was also included to validate the model's specificity and ensure that the identified correlations were not due to random chance or technical artifacts.

Results: Widespread Genetic Sharing

EncoreDNM identified 25 significant pairs of disorders with concordant genetic enrichment. One of the most striking findings was the high genetic correlation between Congenital Heart Disease (CHD) and Cerebral Palsy (CP). This suggests that the same genetic "machinery" involved in building a heart is also critical for brain development, providing a genetic basis for the clinical co-occurrence of these conditions.

Simulation results show that EncoreDNM provides accurate and well-calibrated estimates for model parameters across different scenarios.

In extensive simulations, EncoreDNM demonstrated superior accuracy and calibration compared to existing tools like mTADA. While other methods often produced inflated false-positive rates in small sample sizes, EncoreDNM remained robust. The analysis further revealed that the strongest genetic correlations are driven by Loss-of-Function (LoF) mutations—the most severe type of genetic disruption that completely breaks a gene's protein product. Furthermore, genes shared across multiple disorders were significantly enriched for functions related to chromatin organization, highlighting a common mechanism of transcriptional dysregulation across diverse clinical phenotypes.

The study also identified 26 specific genes that harbor severe mutations across at least four different disorders. A standout was CTNNB1, a gene involved in cell signaling that appeared as a risk factor across a wide spectrum of neurodevelopmental and physical conditions. These shared genetic effects were found to be most prominent in genes highly expressed during fetal development, particularly in brain and heart tissues.

Discussion and Future Directions

EncoreDNM represents a significant advancement in our ability to characterize the shared genetic landscape of complex disorders. By moving beyond a focus on individual significant genes, the framework uncovers widespread pleiotropy that was previously invisible.

However, some limitations remain. The model assumes sample independence between studies; overlapping participants could potentially inflate correlation estimates. Additionally, while EncoreDNM is a powerful tool for quantifying sharing, it is not primarily designed for the discovery of novel risk genes or for determining the direction of causality between linked disorders. Future iterations could involve using these identified correlations to improve gene discovery by "borrowing" statistical power from related disorders, effectively increasing the sensitivity of exome-wide association studies.

Summary updated

Abstract

Exome sequencing on tens of thousands of parent-proband trios has identified numerous deleterious de novo mutations (DNMs) and implicated risk genes for many disorders. Recent studies have suggested shared genes and pathways are enriched for DNMs across multiple disorders. However, existing analytic strategies only focus on genes that reach statistical significance for multiple disorders and require large trio samples in each study. As a result, these methods are not able to characterize the full landscape of genetic sharing due to polygenicity and incomplete penetrance. In this work, we introduce EncoreDNM, a novel statistical framework to quantify shared genetic effects between two disorders characterized by concordant enrichment of DNMs in the exome. EncoreDNM makes use of exome-wide, summary-level DNM data, including genes that do not reach statistical significance in single-disorder analysis, to evaluate the overall and annotation-partitioned genetic sharing between two disorders. Applying EncoreDNM to DNM data of nine disorders, we identified abundant pairwise enrichment correlations, especially in genes intolerant to pathogenic mutations and genes highly expressed in fetal tissues. These results suggest that EncoreDNM improves current analytic approaches and may have broad applications in DNM studies.
BibTeX
@inproceedings{guo2022quantifying,
  author = {Hanmin Guo and Lin Hou and Yu Shi and Sheng Chih Jin and Xue Zeng and Boyang Li and Richard P. Lifton and Martina Brueckner and Hongyu Zhao and Qiongshi Lu},
  title = {{Quantifying concordant genetic effects of de novo mutations on multiple disorders}},
  booktitle = {eLife},
  year = {2022},
  doi = {10.7554/elife.75551},
}
Figure