Identification of novel candidate risk genes for myelomeningocele within the glucose homeostasis/oxidative stress and folate/one-carbon metabolism networks

Abstract

Background: Neural tube defects (NTDs) are the second most common complex birth defect, yet, our understanding of the genetic contribution to their development remains incomplete. Two environmental factors associated with NTDs are Folate and One Carbon Metabolism (FOCM) and Glucose Homeostasis and Oxidative Stress (GHOS). Utilizing next-generation sequencing of a large patient cohort, we identify novel candidate genes in these two networks to provide insights into NTD mechanisms.

Methods: Exome sequencing (ES) was performed in 511 patients, born with myelomeningocele, divided between European American and Mexican American ethnicities. Healthy control data from the Genome Aggregation database were ethnically matched and used as controls. Rare, high fidelity, nonsynonymous predicted damaging missense, nonsense, or canonical splice site variants in independently generated candidate gene lists for FOCM and GHOS were identified. We used a gene-based collapsing approach to quantify mutational burden in case and controls, with the control cohort estimated using cumulative allele frequencies assuming Hardy-Weinberg equilibrium.

Results: We identified 45 of 837 genes in the FOCM network and 22 of 568 genes in the GHOS network as possible NTD risk genes with p < 0.05. No nominally significant risk genes were shared between ethnicities. Using a novel approach to mutational burden we identify 55 novel NTD risk associations.

Conclusions: We provide a means of utilizing large publicly available sequencing datasets as controls for sequencing projects examining rare disease. This approach confirmed existing risk genes for myelomeningocele and identified possible novel risk genes. Lastly, it suggests possible distinct genetic etiologies for this malformation between different ethnicities.

Keywords: Myelomeningocele; exome sequencing; folate; glucose; mutation burden.

FIGURE 1

Variant filtering and annotation workflow with summative case and control variant and gene counts. Top. General workflow and filtering parameters applied to data. Bottom. Table summarizing total qualifying variant and gene counts by ethnicity and in total for cases and controls. Variant counts represent the number of unique qualifying variant loci identified in each ethnicity and network. Gene counts represent the number of genes containing at least one qualifying variant in each network and ethnicity. AMR, American Latino; CADD, combined annotation dependent depletion; dbNSFP, database of non‐synonymous single‐nucleotide variant functional predictions; EA, European Americans; FOCM, Folate and One‐Carbon Metabolism; GATK, Genome Analysis Toolkit; GHOS, Glucose Homeostasis and Oxidative Stress; gnomAD, genome aggregation database; MA, Mexican Americans; NFE, Non‐Finnish Europeans; VCF, variant call format.

FIGURE 2

Number of qualifying variant containing genes by individual. Solid line graphs represent the number of individual cases having the number of genes containing a qualifying variant in the glucose homeostasis and oxidative stress network (GHOS) [blue], folate and one‐carbon metabolism network (FOCM) [orange], and total between the two networks (GHOS + FOCM) [grey] for both Mexican Americans and European Americans. The x‐axis is discontinuous at its upper bound for both graphs as indicated by the hashmarks. Vertical dashed lines represent the 95% population cutoff for each network. Diamond marker represent the population averages for each network.

FIGURE 3

The number of nominally significant genes containing a qualifying variant per individual. Graphs total the number of individual cases having qualifying variants in variable numbers of nominally significant genes of the glucose homeostasis and oxidative stress network (GHOS) [blue], folate and one‐carbon metabolism network (FOCM) [orange], and total between the two networks (GHOS + FOCM) [grey] for both Mexican Americans and European Americans. n provides the population size for each ethnicity.

FIGURE 4

Proportion of qualifying variants and contributing individuals per gene in Folate and One‐Carbon Metabolism network (a) and Glucose Homeostasis and Oxidative Stress (b). Top Rows A and B. Graphs show the total number of qualifying variants (number following “,”) in each nominally significant gene by ethnicity. The central bar graph shows the total number of variants compared between ethnicities (blue bars). Bottom Rows A and B. Similar to top rows, showing the number of individual cases with at least one qualifying variant in the nominally significant genes by ethnicity. The central bar graph shows the total number of affected individuals compared between ethnicities (red bars). Colors are matched between top row and bottom row pairs by ethnicity but not between ethnicities or parts A or B. n provides the population size for each ethnicity.