Whole-exome sequencing reveals known and candidate genes for hearing impairment in Mali

Abstract

Hearing impairment (HI) is the most common neurosensory disorder globally and is reported to be more prevalent in low-income countries. In high-income countries, up to 50% of congenital childhood HI is of genetic origin. However, there are limited genetic data on HI from sub-Saharan African populations. In this study, we investigated the genetic causes of HI in the Malian populations, using whole-exome sequencing. Furthermore, cDNA was transfected into HEK293T cells for localization and expression analysis in a candidate gene. Twenty-four multiplex families were enrolled, 50% (12/24) of which are consanguineous. Clustering methods showed patterns of admixture from non-African sources in some Malian populations. Variants were found in six known nonsyndromic HI (NSHI) genes, four genes that can underlie either syndromic HI (SHI) or NSHI, one SHI gene, and one novel candidate HI gene. Overall, 75% of families (18/24) were solved, and 94.4% (17/18) had variants in known HI genes including MYO15A, CDH23, MYO7A, GJB2, SLC26A4, PJVK, OTOGL, TMC1, CIB2, GAS2, PDCH15, and EYA1. A digenic inheritance (CDH23 and PDCH15) was found in one family. Most variants (59.1%, 13/22) in known HI genes were not previously reported or associated with HI. The UBFD1 candidate HI gene, which was identified in one consanguineous family, is expressed in human inner ear organoids. Cell-based experiments in HEK293T showed that mutants UBFD1 had a lower expression, compared to wild type. We report the profile of known genes and the UBFD1 candidate gene for HI in Mali and emphasize the potential of gene discovery in African populations.

Keywords: Africa; Mali; UBFD1 candidate gene; admixture mapping; hearing impairment.

Graphical abstract

Figure 1

Recruitment of participants, population structure, and variants information (A) Flow diagram of the study. (B) Genes identified in this study. (C) Proportion of novel variants. (D) Type of variants identified in this study. (E and F) Principal-component analysis (PCA) showing that samples cluster within African genomic data when compared to other non-African populations and populations from Gambia (GWD) and Sierra Leone (MSL) as geographically expected (red circle).

Figure 2

Population structure and admixture pattern (A) PCA showing a pattern of genetic diversity consistent with geographical differentiation on PC2 coordinates. (B) PCA7 and PCA8 showing that Fulani individuals from different countries have their own clump due to their genetic affinities, while other studied populations from Mali have their own clump at the other side of the PCA space with Bono, Bambara, and Songhai individuals at the edges of this group. (C) Clustering methods showing patterns of admixture from non-sub-Saharan African sources (blue component) among the studied Fulani populations.

Figure 3

Comparative allele frequencies of common HI variants across global ancestries (A) Allele frequency of the common variants in the top 10 recessive HI genes, showing variability across populations and not found in our cohort. (B) Allele frequency of variants in the most common recessive genes identified in this study across different genetic ethnic groups, showing that these variants were mostly not reported in gnomAD.

Figure 4

Pedigree, genetic and expression data, and three-dimensional (3D) structure of candidate genes (A) The pedigree of family 1 demonstrates an autosomal recessive inheritance pattern with consanguinity, indicated by double horizontal lines between the parents. The affected male is represented by a filled black square, and the affected female is represented by a filled black circle. Unaffected males and females are shown as unfilled squares and circles, respectively. An episode of miscarriage is represented with black dot symbols, asterisks indicate individuals genotyped in this study, and black arrows indicate the probands and capital letters are genotype information. (B) Electropherograms showing the nucleotide change in family 1, indicated by asterisks. (C) Portion of amino acids sequence showing the conservation of the amino acids of interest (red box) across a wide range of species. (D) Uniform manifold approximation and projection 1 plots of UBFD1 gene expression profile in human inner ear organoids. (E) The 3D structures of wild-type (WT) UBFD1 in green and mutant in light blue, showing major structural changes in mutant, including loss of helix and β sheets (red arrow).

Figure 5

Confocal microscopy images and western blot data (A–D) Co-visualization at 10× of the nuclear material (blue) with the GFP (green) signal in WT UBFD1 (A) and mutant (B), and at 40× of WT (C) and mutant (D) showing cytoplasmic localization, with decreased expression in mutant. (E) Western blot analysis showing that both WT and mutant UBFD1 were at the expected size, with lower expression in the mutant.