Burden of Mendelian disorders in a large Middle Eastern biobank

Abstract

Background: Genome sequencing of large biobanks from under-represented ancestries provides a valuable resource for the interrogation of Mendelian disease burden at world population level, complementing small-scale familial studies.

Methods: Here, we interrogate 6045 whole genomes from Qatar-a Middle Eastern population with high consanguinity and understudied mutational burden-enrolled at the national Biobank and phenotyped for 58 clinically-relevant quantitative traits. We examine a curated set of 2648 Mendelian genes from 20 panels, annotating known and novel pathogenic variants and assessing their penetrance and impact on the measured traits.

Results: We find that 62.5% of participants are carriers of at least 1 known pathogenic variant relating to recessive conditions, with homozygosity observed in 1 in 150 subjects (0.6%) for which Peninsular Arabs are particularly enriched versus other ancestries (5.8-fold). On average, 52.3 loss-of-function variants were found per genome, 6.5 of which affect a known Mendelian gene. Several variants annotated in ClinVar/HGMD as pathogenic appeared at intermediate frequencies in this cohort (1-3%), highlighting Arab founder effect, while others have exceedingly high frequencies (> 5%) prompting reconsideration as benign. Furthermore, cumulative gene burden analysis revealed 56 genes having gene carrier frequency > 1/50, including 5 ACMG Tier 3 panel genes which would be candidates for adding to newborn screening in the country. Additionally, leveraging 58 biobank traits, we systematically assess the impact of novel/rare variants on phenotypes and discover 39 candidate large-effect variants associating with extreme quantitative traits. Furthermore, through rare variant burden testing, we discover 13 genes with high mutational load, including 5 with impact on traits relevant to disease conditions, including metabolic disorder and type 2 diabetes, consistent with the high prevalence of these conditions in the region.

Conclusions: This study on the first phase of the growing Qatar Genome Program cohort provides a comprehensive resource from a Middle Eastern population to understand the global mutational burden in Mendelian genes and their impact on traits in seemingly healthy individuals in high consanguinity settings.

Keywords: Arab population; Biobank; Consanguinity; Genome sequencing; Mendelian disorders; Middle East; Pathogenic variants; Qatar; Rare genetic disease.

Fig. 1

Summary of study on the burden of Mendelian disorders in a large Middle Eastern biobank from Qatar. An in-depth analysis of pathogenic variants for Mendelian disorders among 6045 Qatari genomes with 77 biochemical and clinical phenotypes obtained from Qatar biobank. The study consists of two parts: first, investigation of the landscape of known and novel pathogenic variants; second, genome wide rare variant burden analysis. Pathogenic and likely pathogenic variants (P/LP) were defined based on intersecting ClinVar and HGMD. Global AF indicates allele Frequency in international databases including 1000 genomes, gnomAD, and ExAC. Other acronyms are explained as follows: AB, allele balance, AD, allele depth; AF, allele frequency

Fig. 2

Identification and description of pathogenic variants in 6,045 Qatari genomes. a Known pathogenic and likely pathogenic variants (P/LP) were obtained by intersecting variants from the Qatari cohort with 20 curated gene panels, ClinVar and HGMD databases. P/LP were defined as those with classes of “Pathogenic” and/or “Likely pathogenic” in ClinVar and “DM” or “DM?” categories in HGMD. Abbreviations; Cln-PLP, ClinVar pathogenic/likely pathogenic; HGMD-DM+, HGMD class of “DM” and/or “DM?”. b Count of genes per number of P/LP variants. c Allele frequency distribution of P/LP variants indicating most are below 1% (Left). Count of P/LP variants versus number of carrier subjects (Right). d Correlation of allele frequencies of P/PL variants in the Qatari dataset and global databases for variants that are rare in both cohorts (Left) and for those that are common in Qatari dataset and rare in global databases (Right)

Fig. 3

Genes with the most pathogenic alleles and corresponding disease phenotypes and panels. Number of P/LP variants was calculated per gene and adjusted for gene length, showing genes with adjusted counts > 0.3. OMIM phenotype and inheritance mode are shown for each gene, in addition to gene panel membership

Fig. 4

Cumulative gene carrier frequencies based on P/PL variants across various Qatari subpopulations. a Number of genes with carrier frequency > 1/200 for the major Qatari subpopulations (SAS is not included due to small sample size). b Cumulative gene carrier frequency (GCF) among the major Qatari subpopulations highlighting genes with GCF > 2%. Genes marked with “+” indicate those for which an associated phenotype/biomarker is currently included in the Qatari newborn screening program. Genes marked with “$” refer to genes in the ACMG Tier 3 list

Fig. 5

Potential novel pathogenic variants associated with extreme quantitative traits in the study cohort. Shown are 39 variants predicted as being deleterious, for which there are at least two homozygotes and found to be associated with extreme phenotype (at > 95th percentile or < 5th percentile) in 58 quantitative phenotypes in the QBB data. Arrowheads indicate direction of association whether it is > 95th or < 5th percentiles. Highlighted in bold (red color) are two genes with effects supported by the literature as indicated in the main text