Background: Although sickle cell anaemia (SCA) is genetically characterised by a single point mutation, patients can manifest varying degrees of clinical severity due to various genetic modulators that affect the phenotype of this disease. The co-inheritance of alpha-thalassemia (α-thalassemia) has been associated with a milder phenotype in SCA patients (e.g. lower stoke rate), but could also result in the increase of vaso-occlusive (VOC) pain episodes. There is a scarcity of data on the co-inheritance of α-thalassemia and SCA in Cameroon. The present study explored the correlation between α-thalassemia, haematological indices, and clinical events in Cameroonian SCA patients. Materials and Methods: For this cross-sectional study, a full blood count and clinical phenotype profile was collected for 262 Cameroonian individuals. Restriction fragment length polymorphism - polymerase chain reaction (RFLP-PCR) was performed for the molecular diagnosis of SCA and for the study of the β-globin (HBB) gene cluster haplotypes. Multiplex Gap-PCR was performed to investigate the 3.7kb and 4.2kb α-thalassemia gene deletions.

P. falciparum malaria remains one of the leading public health problems worldwide. The global tally of malaria in 2018 was estimated at 228 million cases and 405, 000 deaths worldwide. African countries disproportionately carry the global burden of malaria accounting for 93% and 94% of cases and deaths, respectively. Even though most infected children recover from P. falciparum malaria, a small subset (~1%) of cases progresses to severe disease and death. Over the last decade, several genome-wide association studies (GWASs) have been conducted in diverse malaria endemic populations to understand the natural host protective immunity against severe malaria that can provide clues for the development of new vaccines and therapeutics. However, beyond identifying association variants, conventional GWAS approaches can't inform the underpinning biological functions. To bridge this gap, we applied various contemporary statistical genetic analytic approaches to malaria GWAS datasets of diverse malaria endemic populations. First, we accessed malaria resistance GWAS datasets of three African populations (N=~11,000) including Kenya, Gambia and Malawi from European Genome Phenome Archive (EGA) through MalariaGEN consortium standard data accession procedures. We explored the challenges of GWAS approaches in the genetically diverse Africa populations and figured out how various advanced statistical genetic methods can be implemented to address these challenges. We investigated single nucleotide polymorphism (SNP) heritability (h2 g) of malaria resistance in the Gambian populations and determined appropriate quality (QC) thresholds to accurately estimate the h2 g in our dataset. Second, we estimated h2 g in the three populations and partitioned the h2 g into chromosomes, allele frequencies and annotations using the genetic relationship-matrix restricted maximum likelihood approaches. We further created African specific reference panel from African population datasets obtained from 1000 Genomes Project and African Genome Variation Project dataset and computed linkage disequilibrium (LD). We used LD information obtained from these reference panels to compute cell-type specific and none cell-type specific enrichments for GWAS-summary statistics meta-analyzed across the three populations. Our results showed for the first time that malaria resistance is polygenic trait with h2 g of ~20% and that the causal variants are overrepresented around protein coding regions of the genome. We further showed that the h2 g is disproportionately concentrated on three chromosomes (chr 5, 11 and 20), suggesting cost-effectiveness of targeting these chromosomes in future malaria genomic sequencing studies. Third, we systematically predicted plausible candidate genes and pathways from functional analysis of severe malaria resistance GWAS summary statistics (N = 17,000) meta-analyzed across eleven populations in malaria endemic regions in Africa, Asia and Oceania. We applied positional mapping, expression quantitative trait locus (eQTL), chromatin interaction mapping and gene-based association analyses to identify candidate severe malaria resistance genes. We performed network and pathway analyses to investigate their shared biological functions. We further applied rare variant analysis to raw GWAS datasets of three malaria endemic populations including Kenya, Malawi and Gambia and performed various population genetic structures of the identified genes in the three endemic populations and 20 world-wide ethnics. Our functional mapping analysis identified 57 genes located in the known malaria genomic loci while our gene-based GWAS analysis identified additional 125 genes across the genome. The identified genes were significantly enriched in malaria pathogenic pathways including multiple overlapping pathways in erythrocyte-related functions, blood coagulations, ion channels, adhesion molecules, membrane signaling elements and neuronal systems. Furthermore, our population genetic analysis revealed that the minor allele frequencies (MAF) of the SNPs residing in the identified genes are generally higher in the three malaria endemic populations compared to global populations. Overall, our results suggest that severe malaria resistance trait is attributed to multiple genes that are enriched in pathways linked to severe malaria pathogenesis. This highlights the possibility of harnessing new malaria therapeutics that can simultaneously target multiple malaria protective host molecular pathways. In conclusions, this project showed that malaria resistance trait is mainly a polygenic trait which is influenced by genes and pathways linked to blood stage lifecycle of P. falciparum. These findings constitute the foundations for future experimental studies that can potentially lead to translational medicine including development of new vaccines and therapeutics. However, ‘-omics' studies including those implemented in this study, are limited to single datatype analysis and lack adequate power to explain the complexity of molecular processes and usually lead to identification of correlations than causations. Thus, beyond singe locus analysis, the future direction of malaria resistance requires a paradigm shift from single-omics to multi-stage and multi-dimensional integrative multi-omics studies that combines multiple data types from the human host, the parasite, and the environment. The current biotechnological and statistical advances may eventually lead to the feasibility of systems biology studies and revolutionize malaria research.

Whole-Genome Sequencing (WGS) is ushering a new era in healthcare and research in identifying genetic variation in all populations. However, the African populations are still under-represented. Since African populations are being the most genetically diverse with high heterogeneity rate, we need to benchmark the Whole Genome Sequence (WGS) analysis pipeline to ensure reliable mutation detection. Therefore, it is essential to ensure that all steps of WGS downstream analysis are accurate, mainly the variant calling (VC). Current VC tools may produce falsepositive/negative results; such result may produce misleading conclusions in prioritisation of mutation, clinical relevancy and actionability of genes. With such many VC tools, two questions have arisen. Firstly, which tool has a high rate of sensitivity and precision in low either high coverage African sequences, given they have high genetic diversity and heterogeneity? Secondly, does the improvement of the VC result will advance the accuracy of detecting mutation and incidental finding (actionable genes) in African populations? In this project, a total of 100 DNA sequence samples was simulated (of which every 50 samples mimicked the genetics background of African and European, respectively) at different coverage (high and low). In particular, the sensitivity to discover polymorphisms was done by nine different VC tools. These tools were assessed in term of false positive/negative call rate given the simulated golden variants. Combining our result on sensitivity and positive predictive value (PPV). Lofreq performs best in African population data (sens=0.85, PPV=0.983, F-score=0.91) on high/low coverage data; as a result, we chose Lofreq to perform variant calling, and Gene-based annotation is performed to conduct in-sillico predication of mutation on publicly available data (the African Genome Variation and 1000 Genome Project). In doing so, we have leveraged WGS to examine and validate four of burden diseases in the African content, such as communicable diseases: HIV/AIDS, Malaria, Tuberculosis (TB), and Non-communicable diseases: such as Sickle cell disease, these diseases have uniquely shaped ethnic-specific and continental genomics variation and therefore provides unprecedented opportunities to map disease genes across the African continent. Moreover, the current actionable gene recommended by The American College of Medical Genetics and Genomics (ACMG) in the African population and update on additional African-specific actionable genes. Our result suggests African and African diaspora ethnic groups, particularly Bantu and Khoesan ethnics have gene diversity, high proportion of derived allele at low minor allele frequency (0.0 − 01) and the highest proportion of pathogenic variants within HIV, TB, Malaria, Sickle-Cell disease, while non-African ethnic groups including Latin America, Afro-Asiatic European related ethnic groups have high proportion of pathogenic variants within current actionable gene list. Overall, given the observed highest genetic diversity found in African ethnics and African diaspora related ethnics at these four Africa burden diseases and current actionable gene associated, our results support (1) the use of personalised medicine as beneficial to both African continent and worldwide; (2) a recommendation for African-specific actionable list of genes to further improve African and diaspora healthcare.

Huntington disease (HD) is a devastating neurodegenerative condition characterised by a triad of symptoms: behavioural/psychiatric changes, cognitive decline and movement disorder. The dominantly inherited disease-causing mutation is an expanded trinucleotide (CAG) repeat in the Huntingtin(HTT) gene. Clinical symptoms are believed to be the result of degeneration of specific neuronal populations that are susceptible to the presence of a toxic expanded protein product. The disease is incurable and following the onset of symptoms, is progressively debilitating over 10-20 years and eventually fatal. Although typical epidemiological studies of prevalence are challenging for a genetic disorder such as HD, family studies and various other methods of ascertainment have been used to estimate its occurrence in different populations. Prevalence is therefore known to vary geographically; population-specific haplotypes have been hypothesised to be the basis of this variation between ethnic groups. High prevalence estimates for populations with European ancestry led to the supposition that the HD mutation was introduced to different regions by Europeans. In South Africa, a survey in the 1970s estimated that the prevalence of HD in the white and coloured subpopulations was similar at 2 per 100 000 individuals; while that in the black subpopulation was significantly lower, at less than 0.01 per 100 000 individuals. Molecular genetic analyses have since revealed links between the white and coloured subpopulations which would explain the similarity in prevalence; however, our knowledge of the genetics of HD in the black subpopulation, has been sorely lacking. This study provides, for the first time, a comprehensive analysis of the HTT gene in an African population. An evaluation of the normal distribution of CAG-tract sizes highlighted significant differences between the subpopulations. Haplotype analysis identified population-specific disease-associated haplotypes, confirming distinct origins of the HD mutation in the different subpopulations. In a coloured family with the rare juvenile form of the disease, DNA sequencing revealed no novel variants within the immediate vicinity of the CAG-tract that could be associated with the observed instability. This indicates that genome-wide analyses may be more useful in identifying factors related to repeat instability and future investigations are planned for a cohort of South African patients affected by juvenile onset HD.

Sickle Cell Disease (SCD) is a growing global problem with firm roots in sub-Saharan Africa (SSA) representing over 3/4 of the global burden of the disease. The prevalence of the sickle mutation (HbS) in SSA has been amplified by the partial resistance to Plasmodium falciparum malaria, which is endemic along tropical equatorial Africa. Several genetic variants have since been associated with fetal hemoglobin (HbF), the disease-ameliorating globin protein, including variants at three principal loci; BCL11A, HBS1L-MYB intergenic polymorphisms (HMIP1/2) and the β-globin gene cluster, which together account for 10 - 20% HbF variance in SCD patients. Similarly, numerous signalling pathways have been implicated in the regulation of γ-globin expression, however, a complete understanding of the regulation of HbF remains elusive. The overall aims of this project were: 1a) to investigate the known variants in key HbF-promoting loci such as BCL11A erythroid-specific enhancer, BCL11A, HBS1L-MYB intergenic polymorphism (HMIP1/2), the β-globin gene cluster, as well as the influence of the co-inheritance of 3.7kb alpha globin gene deletion in a cohort of SCD patients from Cameroon; and 1b) to validate novel HbF-promoting loci reported in 2 genome-wide association studies (GWAS) carried out in a population of Sardinians (Italy) and SCD patients from Tanzania and explore the influence of known promoter variants in SAR1 associated with HbF in African American patients amongst Cameroonian SCD patients; 2) to investigate the molecular mechanisms of hydroxyurea (HU)-induced production of HbF using a primary erythroid cell model from hematopoietic stem cells (HSCs) derived from umbilical cord blood and lastly, 3) to investigate the prevalence of SCD-related polymorphisms; β-globin gene haplotype, HbS mutation and malaria-resistance variants in 3 SCD-unaffected (HbAA) cohorts from South Africa, Zimbabwe and Malawi.

Includes bibliographical references (leaves 167-200)

Includes abstract. / Includes bibliographical references. / Spinocerebellar ataxia type 7 (SCA7) is a dominantly-inherited neurodegenerative disease, resulting from a CAG trinucleotide repeat expansion in the ataxin-7 gene. The Ataxin-7 protein is known to play a role in transcriptional regulation through association with cellular histone acetylation complexes, and several studies have highlighted the role of transcriptional dysregulation, caused by the presence of mutant Ataxin-7, in the neuronal dysfunction that precedes the onset of disease symptoms.This study aimed to establish patient-derived cell models of SCA7, for use in the investigation of pathogenesis (with particular reference to transcriptional alterations), and in the evaluation of previously-developed therapies for the disease.The high prevalence of SCA7 in the South African population, as a result of a founder effect, makes this disease particularly amenable to allele-specific RNA interference (RNAi)-based therapy. Thus, this study also evaluated the feasibility of these cell models as a vehicle to test previously-developed RNAi therapeutics, using the alteration of expression of key transcripts as a phenotypic marker. SCA7 patient and control dermal fibroblasts were reprogrammed to pluripotency by retroviral transduction. The resultant induced pluripotent stem cell (iPSC) lines were characterised with respect to endogenous markers of pluripotency, differentiation capacity and transgene silencing. These cells were then subjected to neuronal differentiation, the success of which was confirmed by the expression of early neuronal markers.

Bibliography: leaves 116-130. / Hereditary macular degeneration describes a group of conditions causing macular pathology. Stargardt disease (STGD) is the most common inherited juvenile macular dystrophy characterised by severed reduction of central visual acuity and normal peripheral vision. The ABCA4 (adenosine triphosphate binding cassette transporter) gene is the only gene implicated in the autosomal recessive (ar) form of the STGD phenotype, while one genetic locus and one gene have been shown to be causative of the autosomal dominant form.

Includes bibliographical references / Background: Mental health disorders represent a major public health problem in most countries around the world. In South Africa, the lifetime prevalence of psychiatric disorders is 30.3%, with substance-use disorders and mood disorders being the second and third most prevalent classes of lifetime disorders, respectively. Bipolar disorder (BD) has a lifetime prevalence of 1.4% and alcohol use disorder (AUD) a lifetime prevalence of 30.3%, and they are frequently comorbid. Both of these disorders have a relatively high heritability, yet the exact genetic basis of each remains unknown. Genetic variants within the hypothalamic-pituitary-adrenal (HPA)-axis and glutamatergic pathways have previously been implicated in both phenotypes. The aim of this project was to investigate the aetiology of BD and AUD, using high-throughput genomic technologies, bioinformatics, brain-imaging and environmental measures. An additional aim was to assess the genetic aetiology of BD-AUD comorbidity. Methods: For the genetic analysis underlying BD, a South African 'Afrikaner' family was investigated. Whole-genome sequencing (WGS) and whole-genome linkage analysis was performed for individuals with BD Type I (BDI) and unaffected family members using the Illumina HiSeq2000 and Affymetrix Axiom TM Genome-wide CEU 1 Array, respectively. For the AUD analysis, two groups were investigated; a South African adolescent group comprising 80 individuals with AUD and 80 controls, and a group of 8123 individuals from the Avon Longitudinal Study of Parents and Children (ALSPAC) birth cohort. The South African group of adolescents were genotyped using the Illumina Infinium iSelect custom 6000 BeadChip, childhood trauma data was obtained and brain magnetic resonance images were collected for a subset of this group. Genotype data on HPA-axis genes were obtained from a previous study for the ALSPAC cohort. The fourth group of individuals investigated in this thesis comprised 233 individuals with BD-AUD comorbidity from the Systemic Treatment Enhancement Program for BD (STEP-BD). Genotype data for genes from the glutamatergic and HPA-axis pathways were obtained from a previous study conducted on these individuals. Results: The chromosomal regions 6p25, 10p14-10p15.1, 11q23-11q25, and 13q21-22 scored the highest LOD scores for BD and the most over-represented pathway in the affected family members was the T-cell receptor signalling pathway. In the South African adolescent group, circadian rhythm genes were associated with AUD and childhood trauma predicted alcohol use in adolescence. The gene-imaging analysis identified a SNP in the glutamate receptor, ionotropic, N-methyl D-aspartate 2B (GRIN2B) gene as being associated with brain volume in the left orbitofrontal cortex and posterior cingulate. HPA-axis genes did not show an association with AUD and no significant gene x environment interactions were detected for AUD in the ALSPAC cohort. Single variants in the glutamatergic genes and HPA-axis were not associated with BD-AUD comorbidity. However, from the gene-based analysis, the glutamatergic gene PRKCI was associated with BD-AUD comorbidity. Conclusions: It appears that disruption in immune-related genes may contribute to the development of BD in an Afrikaner family. No significant gene x environment interactions were detected for adolescent AUD. The circadian pathway and childhood trauma may play a role in the development of adolescent AUD. Differential brain volume and BD-AUD comorbidity may be characterised by variation in the glutamatergic pathway. These pathways and the interactions between them should be further investigated in BD and AUD.

Hearing Loss (HL) is one of the highest contributors to disability worldwide. The highest incidence of the disease is seen in developing countries, such as those in subSaharan Africa (SSA). Patients affected with disabling HL are reported to be more than 466 million worldwide. The causes of HL can either be environmental or genetic with each contributing about 50% towards all cases, in many settings. In developing countries, the environment might contribute more due to poor health services and infrastructure available to the population. In the absence of environmental causes, there is a genetic component at play, that is largely unknown in African populations. Up to 70% of HL of genetic origin are non-syndromic (NS). The mode of inheritance is recessive in nearly 77% of non-syndromic HL. Up to date, more than 100 genes have been associated with HL harbouring more than 1000 causative variants. In many populations of European and Asian descent, pathogenic variants in GJB2 (connexin gene 26) and GJB6 (connexin gene 30) are a major contributor to autosomal recessive non-syndromic hearing loss (ARNSHL). Comprehensive hearing health care programs should cover genetic causes by providing molecular testing, and genetic counselling, specifically SSA where genes and mutations causing HL remain largely unknown. The aim of this project was thus to uncover the genetic causes of HL among patients’ cohorts from Cameroon and South Africa. This was addressed by 1) sequencing common variants in the most relevant genes in other populations (GJB2 and GJB6), 2) using a targeted gene panel to resolve HL in 10 multiplex families from Cameroon presenting with ARNSHL and negative for GJB2 and GJB6 mutations screening, 3) screening novel variants found in known genes in a cohort of 82 singleplex HL cases from Cameroon and South Africa, and lastly, 4) using Whole Exome sequencing to explore the two unresolved multiplex cases with and subsequent findings confirmed by functional studies, and also screened in 80 singleplex HL cases. The following findings are reported: GJB6, GJA1 mutations screening and literature review No GJA1 or GJB6 mutation was not found in multiplex and simplex cases of HL in both Cameroonians and South Africans. The review of the literature confirms that the prevalence of GJB2- or GJB6-related NSHL is approximating to zero in most subSaharan African populations. Targeted Exome Sequencing (OtoSCOPE) The targeted genes, panel that included 116 genes, was able to resolve 7 of 9 families (77.8%) which were successfully sequenced, with one family failing to be sequenced. The causative variants identified in the 7 resolved families were : 1) compound heterozygous c.5806_5808delCTC and c.5880_5882delCTT in MYO7A; 2) compound heterozygous c.646T>A (p.Phe216Ile) and c.38G>A (p.Arg13His) in LOXHD1; 3) homozygous c.766-2A>G in OTOF; 4) a deletion and a complex copy number variation in STRC; 5) compound heterozygous c.1678G>A (p.Asp560Asn) and c.2007C>A(p.Asp669Glu) in SLC26A4; 6) Homozygous c.1996C>T(p.Arg666Stop) in MYO7A; 7) compound heterozygous c.6399C>A(p.Asp2133Glu) and c.2000T>C (p.Met667Thr) in CDH23. Five out of 12 variants were novel. Screening of these causative variants in known genes, in 82 singleplex HL cases from Cameroon and South Africa was unable to resolve any of the cases: the variants were in either heterozygous in low frequency or absent. Bioinformatic pathways exploration of SNP data of known HL genes revealed an extensive network within the HL genes, with 10 identified as important nodes, including MYO7A. Most HL genes were found to be involved in two biological processes which were sensory perception of mechanical stimulus (GO: 0050954, p= 1.430e-8) and sound (GO: 0007605, p = 1.246e-8). The molecular functions of variants found within these genes were found to mostly fall within the binding (GO: 0005488) and/or structural molecule activity (GO: 0005198). Whole Exome sequencing Whole exome sequencing was performed on four of the nine multiplex families: the two families that were unresolved by targeted panel sequencing, and two previously resolved families that were used as positive controls for the variant annotation and filtering pipeline. The results were the resolution of 3/4 families, including the two- positive control. The previously unresolved “family 8” was found to harbour a novel variant within the GRXCR2 gene, a gene only associated with HL once before. The c.251delC variant was revealed through in silico studies to cause a premature stop codon at position 116 due to its frameshift effect. The screening of this variant in our cohort of 80 singleplex cases revealed one other unrelated HL patient harbouring this causative variant. Due to the limited literature on the gene and its protein, in silico studies were used to show the predicted secondary structure folding of the protein as well as potential protein binding regions. Analysis showed that the predicted loss of a stable region of the protein as well as that of a putative binding domain could explain the pathogenic nature of the variant. In vitro studies showed that the variant hindered the detection of the protein by way of a DDK tag downstream in the plasmid. Additionally, GFP-Tagged GRXCR2 showed altered expression pattern in the variant when compared to the wildtype. In summary, our data has revealed the efficacy of using next generation sequencing tools in resolving HL among sub-Saharan African patients as opposed to the single candidate gene approach. In our quest, we have employed two widely used strategies, targeted panel and whole exome sequencing (WES), both of which have had great successes in various populations. The targeted approach was able to resolve 77.8% of our families but did not detect variants for two of the families revealing the presence of other variants harboured in rarely associated gene not captured or included on the panel. This prompted for the use of a more comprehensive approach such as WES. These results corroborated with those of two families previously resolved by targeted exome sequencing. Additionally, one of the previously unresolved family was now resolved. This showed that WES was sensitive enough to detect variants in known HL genes but comprehensive enough to detect variants in other regions of the exome which have not been associated with HL or rarely associated with HL. The benefit of WES also extends to the contribution of exomic data from patients of African descent as there is an underrepresentation of this group in exome repositories as well as genomic or SNP databases. To the best of our knowledge, this is the first study to use WES to resolve HL in patients of African descent. The other benefit of such a venture is the use of this data not only for patients in SSA but also those in the diaspora. In conclusion, we have successfully demonstrated the feasibility of using NGS tools in identifying causative variants in HL patients in SSA. Additionally, we have shown that WES is a more suitable approach to trying to resolve HL in Africa. Therefore, the data strongly support that genetic studies on families segregating HL in SSA could be the next frontier of HL genetic research, of global importance through discovering novel variants in known genes, and potentially novel genes. These studies will improve HL genetic diagnosis, retrospective counselling and testing, prevention and care including future prediction of treatment outcomes in sub-Saharan Africans, and in people of African descent.