African populations are vastly underrepresented in genetic studies despite having the most genetic variation globally and facing wide-ranging environmental exposures. Most of these studies have been conducted in populations of European (EUR) ancestry using GWAS arrays that represent the genetic variation in these populations. Thus, the prediction accuracy of polygenic risk scores (PRS) derived from EUR ancestry populations is less accurate in populations of non-European ancestry, and least accurate in African (AFR) ancestry populations. The extent to which PRS prediction accuracy varies within AFR ancestry populations has not, however, been previously investigated. This study had two aims: the first was to investigate the contribution of common variants to the risk of schizophrenia in the South African Xhosa (SAX) population through genome-wide association study (GWAS) analysis, and to determine if PRS derived from EUR and East Asian (EAS) ancestry populations from the Psychiatric Genomics Consortium (PGC) Schizophrenia Working Group were generalizable to SAX. The second aim was to assess the generalizability of PRS for non-psychiatric phenotypes that were derived from EUR ancestry individuals from the UK Biobank (UKB, n = ~350,000) in the Uganda General Population Cohort (GPC, n = 4,778) and the South African Drakenstein Child Health Study (DHCS, n = 638). To address the first aim, a GWAS was conducted in 2,086 Xhosa individuals from South Africa with and without schizophrenia (ncases = 1,038; ncontrols = 1,048) using a custom-designed Affymetrix GWAS array designed to capture variation in the Xhosa population. The schizophrenia GWAS in SAX yielded one SNP (rs35172303 ; P = 4.74e-08, OR = 0.6004, 95%CI:[0.499,0.721]) in ZFP3 that met genome-wide significance. The association of variants in ZFP3 from the schizophrenia GWAS is consistent with those from an earlier exomesequence study in SAX undertaken by colleagues, but this gene has not previously been associated with schizophrenia in large-scale schizophrenia GWAS of predominantly EUR ancestry. After characterizing the genetic architecture of schizophrenia in SAX, it was found that the heritability was enriched across functional categories involved in the regulation of gene expression. Then, the accuracy of PRS derived from PGC Schizophrenia Working Group from both EUR and EAS ancestries in predicting schizophrenia in SAX was quantified. There was low PRS prediction accuracy using PGC-derived summary statistics in SAX (PGC-EUR: max R2 = 0.0057, P = 0.008; PGC-EAS: max R2 = 0.0059, P = 0.007). These findings are consistent with previous findings that showed that PRS predication accuracy is low when discovery and target cohorts come from different ancestral backgrounds. For the second aim, PRS prediction accuracy was quantified in simulations using data from the African Genome Variation project (AGVP) to represent continental AFR diversity. Samples were categorised by geographical region into West, East and South Africa cohorts. Each cohort was divided into a discovery and target datasets. The West and East African discovery data was used to predict the simulated phenotype in the three target cohorts. Using UKB EUR ancestry individuals, PRS prediction accuracy was assessed for 34 anthropometric and blood panel traits in the Uganda GPC, and then meta-analysed UKB with PAGE (Population Architecture using Genomics and Epidemiology, comprising about 50,000 Latino/Hispanic and African-American individuals) and BBJ (Biobank Japan, n = ~162,000) to assess how the inclusion of diverse sample impacts PRS prediction accuracy. Simulations were limited by sample size but showed that PRS prediction accuracy was highest when the discovery and target cohorts were matched by African region, and for phenotypes with the sparsest genetic architecture. Using empirical data from UKB and the Uganda GPC, a low prediction accuracy was observed across all 34 quantitative traits in GPC when using GWAS data from UKB. There was differential prediction accuracy across AFR ancestry groups within UKB, i.e. the prediction accuracy was highest for the Ethiopian and admixed populations, and lowest for southern African populations. When comparing PRS prediction accuracy of East African individuals from the UKB to that of individuals from GPC, the prediction accuracy was lowest in the Ugandan GPC population, indicating that the difference in environments between the two groups may be contributing to the difference in PRS accuracy. Moreover, the cross-ancestry meta-analyses showed that the inclusion of diverse samples in large scale studies improves PRS prediction accuracy, most especially for phenotypes with population-enriched variants. It was demonstrated for the first time in this thesis that EUR ancestry-derived PRS prediction accuracy varied within continental AFR ancestry groups, and tracks with population history and the evolution of humans. The higher prediction accuracy observed in Ethiopians can be explained by their genetic proximity to Europeans as a result of the back to Africa migration, whereas the southern African populations (including SAX) are more proximal to the ancestral populations that never left the continent. It is therefore imperative to not only include more African samples in future large-scale studies, but to have samples that adequately represent the genetic and environmental diversity on the African continent.

Sickle cell disease (SCD) is a common blood disease caused by a single nucleotide substitution (c.20T>A, p.Glu6Val) in the beta globin gene on chromosome 11. The prevalence of the disease is high throughout large areas in sub-Saharan Africa, the Mediterranean basin, the Middle East, and India due to the level of protection that the sickle cell trait, provides against severe malaria. Approximately 300,000 infants are born per year with sickle cell anemia, which is defined as homozygosity for the sickle hemoglobin (HbS). The majority (nearly 75%) of these births occur in sub-Saharan Africa, particularly in two countries: Nigeria, and the Democratic Republic of the Congo where there are poorly resourced healthcare systems. Early diagnosis, penicillin prophylaxis, blood transfusions, hydroxyurea, and hematopoietic stem-cell transplantation can dramatically improve survival and quality of life for patients with SCD. However, our understanding of the role of genetic and clinical factors in explaining the complex phenotypic diversity of this disease is still limited. Early prediction of the severity, and patients' responses to specific therapeutics of SCD could lead to more precise treatment and management. Beyond well-known modifiers of disease severity, such as fetal hemoglobin (HbF) levels and αthalassemia, other genetic variants might influence specific sub-phenotypes. New treatments and management strategies accounting for these genetic and nongenetic factors could substantially and rapidly improve the quality of life and reduce health care costs for patients with SCD. Patients with SCD are subjected to long term administration of drugs and there is a limited data on pharmacogenomics of SCD therapeutics. Vaso-occlusive crisis (VOC) are the main clinical events of SCD and are associated with recurrent and long-term use of antalgics/opioids and HU. This project aimed to investigate the clinical and genetic predictors of painful vaso-occlusive crisis (VOC) among SCD Cameroon patients by exploring pharmacokinetic determinants of treatment responses as well as post-transcriptional signatures triggered by hydroxyurea treatment, particularly, miRNA expression. SCD patients were recruited from Yaounde Central Hospital and Laquintinie Hospital in Douala (Wonkam et al., 2018, Mnika et al., 2019 (b)), and recent migrants SCD patients from the DRC, recruited at the Haematology Clinic, Groote Schuur Hospital in Cape Town, South Africa (Mnika et al., 2019 (a) and Mnika et al., 2019 (b)). Sociodemographic and clinical data were collected by means of a structured questionnaire. Patients' medical records were reviewed to extract their clinical features over the past 3 years. Specifically, the occurrences of VOC, hematological parameters, hospital outpatient visits, hospitalisation, overt strokes, blood transfusions, and administration of hydroxyurea were recorded. Height, weight, body mass index (BMI), systolic and diastolic blood pressures (SBP and DBP) were measured. Detailed descriptions of patients and sampling methods used in the Cameroonian patients have been reported previously (Wonkam et al., 2018 Mnika et al., 2019 (a) and Mnika et al., 2019 (b)). For the purpose of comparing frequencies of variants, ethnically matched Cameroonian controls were randomly recruited from apparently healthy blood donors in Yaounde for participation in the study. All blood samples were collected for genomic characterisation and analysis. DNA was extracted from peripheral blood, following instructions on the available commercial kit [QIAamp DNA Blood Maxi Kit ® (Qiagen, United States)]. Genotyping (TaqMan and MassArray) was performed for 40 variants in 17 pain-related genes, three fetal haemoglobin (HbF)-promoting loci, two kidney dysfunction-related genes, and HBA1/HBA2 genes for 436 patients. A subset of these samples was also genotyped to analyse 32 core and 267 extended pharmacogenes using commercially available PharmacoScan® platform for characterisation of pharmacokinetic determinant of response. We also compared the pharmacogenes variants from these African groups, to data extracted from the 1000 genomes Project. Moreover, association studies were carried out on pharmacogenes variants with SCD clinical variability. Additionally, protein-protein interaction (PPI) network and enriched biological processes and pathways were investigated. For association studies, statistical models using regression frameworks to analyse 40 variants were performed in R®. For miRNA expression, total RNA was isolated using the miRNeasy kit according to protocol of the Manufacturer (QIAGEN, Hilden, Germany); and sequenced by the Genomic and RNA Profiling Core at Baylor College of Medicine, United States, using the NanoString Platform (NanoString Technologies, Inc., Seattle, WA, United States), according to manufacturer's instructions. Genes with statistically significant changes in expression were analysed using the significance analyses of microarrays (SAM) tools. Female sex, body mass index, Hb/HbF, blood transfusions, leucocytosis and consultation or hospitalisation rates significantly correlated with VOC. Three painrelated gene variants correlated with VOC (CACNA2D3-rs6777055, P = 0·025; DRD2- rs4274224, P = 0·037; KCNS1-rs734784, P= 0·01). Five pain-related gene variants correlated with hospitalization/consultation rates (COMT-rs6269, P = 0·027; FAAHrs4141964, P = 0·003; OPRM1- rs1799971, P = 0·031; ADRB2-rs1042713; P < 0·001; UGT2B7-rs7438135, P = 0·037). The 3·7 kb HBA1/HBA2 deletion correlated with increased VOC (P = 0·002). HbF-promoting loci variants correlated with decreased hospitalisation (BCL11A-rs4671393, P = 0·026; HBS1L-MYB-rs28384513, P = 0·01). APOL1 G1/G2 correlated with increased hospitalisation (P = 0·048). A commercial genotyping array platform (PharmacoScan®) with 4627 markers located in 1191 genes was used to investigate 299 pharmacogenes (32 ADME core and 267 extended pharmacogenes). Based on the PharmacoScan analyses, no statistically significant differences in allele frequencies were detected between SCD cases and controls from Cameroon. A principal component analysis (PCA) revealed that Cameroonians' data clustered with other Africans, but this population is significantly distinct from American, European and Asian populations data. Variant allele frequencies in 21/32 core pharmacogenes were significantly different between the two SCD groups (Cameroon vs. Congo). No correlation between clinical variability and variants in the core genes was detected for both populations under study. An association study of the core and extended PharmacoScan variants to VOC identified statistically significant associations between two single nucleotide polymorphisms (SNPs) to VOC after correction of multiple testing. These two SNPs mapped to 50 genes, with two SNPs located in core pharmacogenes (SLCO4A1- rs118042746, p=1.21e-07; UGT1A10, UGT1A8- rs10176426, p=1.22e-07). Functional enrichment analyses revealed that these 50 genes are involved in three biological processes and four pathways relevant to SCD pathophysiology, including xenobiotic glucuronidation (GO:0052697, p = 2.3e-03), and drug metabolism - other enzymes (p = 2.1e-02). Further analyses of the 50 genes, identified key genes in human proteinprotein networks: NTSR1, LRMDA, SMAD SMAD4 and CDH2. These four genes also interacted with three core pharmacogenes associated with VOC: UGT1A8, UGT1A10 and SLCO4A1. We found 22/798 miRNAs to be differentially expressed under HU treatment, with the majority (13/22) being functionally associated with HbF-regulatory genes, including BCL11A (miR-148b-3p, miR-32-5p, miR-340-5p, miR-29c-3p), MYB (miR-105-5p), KLF-3 (miR-106b-5), and SP1 (miR-29b-3p, miR-625-5p, miR-324-5p, miR-125a-5p, miR-99b-5p, miR-374b-5p, miR-145-5p). The present thesis started by highlighting the scarcity of studies investigating variable responses to pain in SCD patients and then proceeded to addressing this research gap. To our knowledge this is the first body of from Africa to provide evidence supporting the possible development of a genetic risk model for pain in SCD. This is also the first body of work to report an association between these two SNPs and VOC in core and extended pharmacogenes. Our data reveals that the commercial pharmacogenes arrays investigated might need additional evidence for appropriateness among Africans. Therefore, it advocates the need to invest in research exploring population-specific arrays, drug design, targeting, and efficacy, for improved clinical management of patients of African descent. Previous studies have investigated various mechanisms to understand the genomic variations affecting responses to HU, but full understanding of the variable HU-mediated HbF production among individuals affected by SCD remains elusive. The present study showed that mechanisms of HbF production in response to HU, could particularly be mediated through miRNA regulation. The data reveals some alternative perspectives and routes towards identifying new therapeutic targets and approaches for SCD. However, this study needs to be replicated in larger samples in multiple African populations.

Activity-based training (ABT) represents the current standard of care in neurological rehabilitation centers around the world. However, innovative rehabilitation techniques have been developed including robotic locomotor training (RLT). The conceptual basis for RLT initially appeared promising; a rehabilitation modality that removes the need for intensive assistance from therapists, whilst facilitating safe and effective over-ground ambulation. However, small sample sizes and a lack of homogeneity across studies have resulted in an underpowered evidence base supporting the efficacy of RLT for SCI rehabilitation. Thus, this randomized control pilot study aimed to investigate the effects of RLT compared to ABT on functional capacity, secondary complications, and psychological well-being in people with SCI after 24-weeks of rehabilitation. Participants with chronic, traumatic motor incomplete SCI were randomized into two intervention groups: RLT (n = 8) and ABT (n = 8) groups. RLT involved solely walking in the Ekso bionic suit. ABT involved a variety of resistance, cardiovascular and flexibility training combined with regular weight-bearing in the standing position. Outcome measures, including functional strength, ambulatory function, pain, spasticity, bladder/bowel, bone density, body composition, quality of life (QoL) and depression were tested at baseline, 6, 12 and 24-weeks of the intervention. There were no significant differences between the intervention groups for lower or upper extremity motor scores (UEMS effect size (ES) = 0.30; LEMS ES = 0.07), back strength (ES = 0.14) and abdominal strength (ES = 0.13) after training. However, both groups showed a significant increase of 2.00 points in UEMS and a significant increase in abdominal strength from pre- to post intervention. Only the RLT group showed a significant change in LEMS, with a mean increase of 3.00 [0.00; 16.5] points over time. Distance walked in the Functional Ambulatory Inventory (SCI-FAI) increased significantly (p = 0.02) over time only for the RLT group. Therefore, the RLT showed promising evidence for potentially inducing functional strength changes and improvements in ambulatory function after 24 weeks of training. There was some evidence to support RLT to induce bowel improvements in individuals with SCI and both interventions appeared to reduce urinary incontinence and improve bladder function (p = 0.04). Total spasticity and pain intensity were similar between groups (p = 0.25; p = 0.96). However, pain interference ratings significantly increased from pre-post intervention for both groups (p = 0.05). RLT prevented the progressive decline of bone mineral density usually occurring in the SCI population, as hip BMD was maintained during RLT; however, it was significantly reduced (p = 0.04) during ABT, with a mean reduction of 0.06 [-0.34, 0.22] g/cm2 (5%) from pre to post intervention. No change in leg fat-free soft tissue mass (FFSTM) occurred between groups or over time (p = 0.32), however, there was a significant 7% increase in arm FFSTM over time for both groups (p < 0.01). The ABT group was more effective (ES = 1.02) in reducing central and peripheral adiposity, with a significant decrease in visceral adipose tissue (VAT) (p = 0.04) and gynoid FM (p = 0.01) over time. Both groups reported increased QoL and decreased depression ratings over time, with the RLT group having a significant change in the general life and physical health domains, p = 0.03, respectively. This pilot trial offers promising evidence for the effectiveness of RLT for improving functional and ambulatory capacity, reducing secondary complications, and potentially improving QoL in people with incomplete SCI. Thus, this dissertation adds substantial weight to the lacking evidence base on the effects of RLT, by incorporating a large homogenous sample, comprehensive testing procedures and an extended intervention period within South Africa.

Background: Several variants have been associated with the risk of musculoskeletal soft tissue injuries, suggesting a role for genetics in the aetiology of chronic Achilles tendinopathy (AT) and anterior cruciate ligament (ACL) ruptures. Genetic risk modifiers have primarily been identified using a hypothesis driven candidate gene approach. However, the ability to identify all risk-conferring variants using this approach alone is limited. Therefore, the primary aim of this thesis was to further define the molecular signatures of musculoskeletal soft tissue injuries mapping to specific genes. The genes encoding the tenascin-C glycoprotein (TNC, chromosome 9), the α1 chain of type XXVII collagen (COL27A1, chromosome 9), matrix metalloproteinase 3 (MMP3, chromosome 11) and the α1 chain of type I collagen (COL1A1, chromosome 17) were previously associated with the risk of injury and were therefore prioritised for further interrogation. Variants within these regions, which had either been previously associated with injury risk or prioritised from the list of new candidates identified by whole exome sequencing (WES) through the application of a customised tiered filtering strategy, were genotyped in several self-identified white AT and ACL rupture cohorts. The second aim of this study was to determine whether the observed risk-associated variants in the self-identified white cohorts were similar to those underpinning injury in the ancestrally admixed South African Coloured cohort, using ACL ruptures as the phenotypic model. The specific objectives of this thesis were: • To select well-phenotyped participants to be sequenced using an extended whole exome sequencing platform • The development and application of a reusable bioinformatics analyses pipeline involving a customised, tiered filtering strategy to select candidates for interrogation from the list of variants identified by WES. The TNC, COL27A1, MMP3 and COL1A1 genes were prioritised for further interrogation using this approach. • To test the association between the selected candidate variants and the risk of chronic AT and ACL ruptures using a case-control genetic association study design. The candidates selected from the list of variants identified by WES included: TNC rs1061494 (T>C), rs2104772 (T>A) and rs1061495 (T>C) and COL27A1 rs2567706 (A>G), rs2241671 (G>A) and rs2567705 (A>T). In addition, several variants previously associated with the risk of injury including TNC rs1138545 (C>T), MMP3 rs3025058 (5A>6A), rs679620 (G>A), rs591058 (C>T) and rs650108 (G>A) and COL1A1 rs1107946 (G>T) and rs1800012 (G>T), were also prioritised for investigation in additional injury cohorts. • To functionally annotate the prioritised variants using a host of in silico bioinformatic analyses tools. Methods Whole exome sequencing and data processing: Ten asymptomatic controls and ten chronic AT cases were selected for sequencing. Controls were older than 47 years of age, were physically active and had not reported any previous tendon or ligament injuries. Cases were younger than 35 years of age, had suffered chronic, bilateral Achilles tendinopathy and/or reported several Achilles tendon injuries. Paired-end WES was performed on the Illumina HiSeq 2000/2500 platform at 30X coverage. A customised, tiered filtering strategy was developed to screen for candidate variants. All candidate variants were confirmed using Sanger Sequencing and genotyped, together with the other prioritised variants in the larger injury cohorts. Case-control genetic association studies Achilles tendon injury cohorts: Three cohorts were independently recruited from South Africa (SA), Australia (AUS) and the United Kingdom (UK). The South African White (SAW)- Achilles tendon injury cohort consisted of 165 controls (SAW-CONAT), 123 cases with chronic Achilles tendinopathy (SAW-TEN) and 47 cases with acute Achilles tendon ruptures (SAWRUP). The UK-Achilles tendon injury cohort consisted of 130 controls (UK-CON), 87 cases with chronic Achilles tendinopathy (UK-TEN) and 35 cases with acute Achilles tendon rupture (UKRUP). The AUS-Achilles tendon injury cohort included 210 controls (AUS-CON) and 85 cases with chronic Achilles tendinopathy (AUS-TEN). Anterior cruciate ligament rupture cohorts: The first ACL rupture cohort consisted of 232 control participants (SAW-CONACL) and 234 cases with surgically diagnosed ACL ruptures (SAW-ACL), of which 135 were reportedly non-contact in mechanism (SAW-NON). All participants in this group were self-identified to be of South African White ancestry. The participants in the second South African ACL rupture cohort were of mixed ancestry and self-identified as being South African Coloured (SAC). This group consisted of 100 controls (SAC-CON) and 97 participants with surgically diagnosed ACL ruptures (SAC-ACL), of which 50 were reportedly non-contact in mechanism (SAC-NON). The TNC and COL27A1 genomic intervals were explored through the TNC rs1061494, rs1138545, rs2104772 and rs1061495 variants and the COL27A1 rs2567706, rs2241671 and rs2567705 variants in the SAW- and UK-Achilles tendon injury cohorts, in addition to the SAWand SAC-ACL rupture cohorts. The MMP3 locus was explored using the previous riskassociated rs3025058, rs679620, rs591058 and rs650108 variants in the AUS-Achilles tendon injury and SAW-ACL rupture cohorts. Chapter 5 explored the COL1A1 locus using the previous risk-associated COL1A1 rs1107946 and rs1800012 variants in the SAW- and UK-Achilles tendon injury cohorts, in addition to the SAW-ACL rupture and SAC-ACL rupture cohorts. Statistical analyses were performed using the R programming environment, with statistical significance set at PC) variant is predicted to overlap the sequence motifs of the muscle initiator nuclear protein, members of the myogenic family of transcription factors and RNA polymerase II subunit A. Furthermore, the rs1061494 variant demonstrated marked differences in its predicted pre-mRNA structure. The other TNC variant associated with injury risk, rs2104772 (T>A), was predicted to be deleterious by two independent annotation tools. The investigated COL27A1 variants were suggested to interact with several predicted enhancers of cellular function. However, this gene is still relatively uncharacterised in musculoskeletal soft tissue injuries, and therefore, these variants will require further interrogation. The 8kb MMP3 genomic interval demonstrated high levels of linkage disequilibrium. Furthermore, MMP3 was predicted to interact with the MMP12 gene mediated by chromatin looping. The COL1A1 rs1800012 (G>T) variant overlaps the recognition sequence of the Sp1 transcription factor in intron 1. This variant is also proposed to interact with the functional promoter variants, rs1107946 (G>T) and rs11327935 (indel/T). This interaction is suggested to be mediated by chromatin looping. Furthermore, the rare rs1800012 T allele is predicted to result in a looser mRNA conformation immediately surrounding the variant within the pre-mRNA sequence compared to the ancestral G allele. Conclusion: These results provide proof of concept for the use of WES and a customised tiered filtering strategy to identify and prioritise variants for further interrogation using traditional molecular techniques. This approach, utilising previous research to guide a targeted analysis of a WES dataset has highlighted the potential risk modifying effects of several new variants in the TNC and COL27A1 genes. Furthermore, haplotype analysis has implicated several signatures encompassing variants previously associated with risk of injury in the four investigated genes. Although no new candidate variants within the MMP3 and COL1A1 genes were independently associated with risk of injury, unique allele combinations were observed to co-segregate with an altered injury risk profile. Therefore, this study has genetically characterised several previously implicated loci and highlighted new sequence signatures, which may potentially contribute to the susceptibility of musculoskeletal soft tissue injuries. The next step would be to explore the functional significance of these sequence signatures in vitro; a process that would help further characterise the biological mechanisms underpinning the observed risk associations.

Background: More than century, malaria is qualified as a mortal infectious disease, worldwide causing high morbidity and mortality. The World Health Organization (WHO) has shown that, Distribution of Malaria in Africa takes a major part, it's accounting for 95% (about 229 million) and 67% (about 274000) of reported cases and death respectively. One of solutions for reducing this threat is to find drugs or to develop vaccines which can resist and adapt to populations. Unfortunately, despite several efforts, malaria parasites are still developing resistance to the frontline antimalarials. Objectives: Our aim in this project is to conduct a systematic Meta-analysis and various functional analysis across three study populations in Africa ( Kenya, Malawi and Gambia ). Method and Materials: Our first analysis is directed to the Genome Wide Association Study (GWAS) of three study populations (Kenya, Malawi and Gambia) using the Emmax tool to identify the genetic variants associated with severe malaria. We then conducted GWAS based meta-analysis on the summary statistics from the three studies using Metasoft and Metal. Further, we implemented Functional GWAS (FGWAS) to re-weight the GWAS meta-analysis using functional genomic information software (fgwas-tool). Using results from fgwas-tool, we performed biological interpretation using Functional Mapping (FUMA) tool. We mapped the significant SNPs to the genes, and elucidated their functions and their associated cell types. We then performed pathway analysis and enrichment analysis of the genes using Genemania and Enrichr. Additionally, we performed a polygenic risk score for individuals in each study population using PRSice, and evaluated the level of risk exposure for each individual based on the best predictive threshold. Finally, we filtered the rare variants from each study, and performed SKAT analysis to aggregate the effect of the rare variants Results: We identified 29 significant SNPs (14 replicates and 15 novels) reweighted from FGWAS based on GWAS Meta-Analysis. The SNPs mapped to 15 genes (HBB, HBD, ATP2B4, ABO, CBLB, EYA2, HERPUDI, IQCJ, MPP7, NAVI, NUP210, SAMD5 , TCERG1L ,TMEM229B, C4orf19) at gene level. Five of these genes (HBB, HBD, ATP2B4, ABO, CBLB) had been reported by different studies to be associated with malaria. In the PRS analysis we have shown the best prediction based on the best threshold estimated of each population. We found best-fit prediction best-fit PRS for Gambia is 0.00443458 at PT = 0.00165005, for Kenya is 8.4666e-158 at PT= 1 and for Malawi is 1.5151e-55 at PT = 1 predict the risk of an infectious disease like severe malaria. However, the prediction rate is very low and may fail to distinguish the cases from the controls. Conclusion: The functional analysis based on fgwas result have shown that 5 genes (ATP2B4, ABO, HBD, HBB, CBLB) are highly associated to malaria across these 3 studies populations (Gambia, Malawi and Kenya) and 10 candidate novel genes, including high number of mutations in the gene C4orf19 which will constitute one of the future major studies. Also, we have shown the best prediction based on the best threshold estimated of each population. The results have shown that the prediction rate is very low and may fail to distinguish the cases from the controls.

Birth registration data forms part of vital statistics. It is the right of the child to be registered immediately after birth and to acquire an identity and a nationality as stipulated by the Convention on the Rights of the Child (UNICEF 1989). The assessment of birth registration is important to help authorities in the processes of planning and decision making. This study investigates birth registration data at a district level with the aim to establish the data's usefulness in determining reliable completeness of births and estimating total fertility rate (TFR). The study further analyses the spatial relationships between respective districts to each other based on levels of birth completeness and total fertility rate. The Geographical Information Systems (GIS) technique of the Global Moran's Index spatial autocorrelation is used to examine the spatial distribution of completeness and TFR. The Indirect method of relational Gompertz model is used to calculate robust estimates of actual births that occurred in the twelve-month period before 2011 Census and 2016 Community Survey, respectively. Then, the Hauer and Schmertmann (2020) method of determining fertility was used to validate results from the Gompertz model. The study establishes that there was an improvement in the promptness of birth registration between 2011 and 2016, highlighted by an 82% completeness in 2011 that increased to 85% in 2016 for births that were registered within the same year of occurrence. This is evidence that mothers are registering births at younger ages than before. The TFR decreased from 2.55 in 2011 to 2.28 in 2016. Apart from that, the study illustrated that districts with higher completeness levels tend to be in major urban agglomerations. However, no spatial relationship could be established meaning that the neighbouring districts do not follow any pattern when compared to each other. It was also noted that districts with low fertility are clustered near major cities. Although there are issues with data at lower levels of disaggregation such as districts, it has been shown that the use of robust methods produces results that help to give meaningful insights of birth registration data.

Background Hearing impairment (HI) is the most common sensory disability and occurs in about 1 per 1000 live births in high-income countries, with a much higher incidence of up to 6 per 1000 live births in sub-Saharan Africa (SSA). HI can be due to environmental or genetic causes, and in many cases, it is not possible to establish a definite aetiology. Hereditary HI contributes to 30% to 50% of HI cases in SSA. Hereditary HI can be syndromic or non-syndromic, depending on whether it is associated with additional abnormalities in other organs or not. Non-syndromic HI (NSHI) accounts for 70% of hereditary hearing loss, and is genetically highly heterogeneous, with approximately 170 loci and 121 genes identified to date. Studies in European and Asian populations have identified pathogenic variants in GJB2 (MIM: 121011), and GJB6 (MIM: 604418) genes as the major contributors to autosomal recessive NSHI (ARNSHI). The genetic aetiology of HI in Cameroon is unclear, as previous studies have found no contribution of GJB2 and GJB6 genes to NSHI in Cameroon. However, patients included in those studies consisted of both familial and isolated cases, therefore, underlying environmental/multifactorial causes in some cases cannot be excluded (especially for the isolated cases). Six loci for X-linked HI have been described to date, including DFNX3 (Xp21.2), where DMD is located. Variants in DMD in humans are known to be responsible for Duchenne muscular dystrophy (DMD; MIM: 310200), and Becker muscular dystrophy (BMD; MIM: 300376), an Xlinked recessive disorder. Previous studies have demonstrated that mdx mice, (an animal knockout model for DMD), have an increased threshold for hearing when compared to wildtype mice. However, the contribution of DMD to HI in humans has not been extensively studied. Besides, most of the previous studies on DMD were conducted in Caucasians, Asians, and Arabs; therefore, little is known about the features of this condition in Africans. Parents of children with HI tend to face challenges of parenting especially in terms of communication and social interaction. In Africa, parent's perceived causes of deafness vary from environmental factors to mysterious (“evil forces”) or superstitious beliefs. Also, the attitude of the society towards people with HI does not encourage their participation and involvement in the community, as they face overt discrimination. Aim and methods The aim of this project was to examine the genetic aetiologies of HI in the Cameroonian population, and undercover the challenges faced by persons with HI in Cameroon and their understanding of the causes of HI. This was addressed by 1) Establishing the current status of knowledge on HI in Africa (in terms of prevalence, aetiologies, and genetics aspects) with a particular focus on Cameroon, and assessing the contribution of connexin genes to HI in humans at a global level, through systematic literature reviews; 2) Revisiting the contribution of GJB2 and GJB6 genes to NSHI in 29 multiplex Cameroonian families with NSHI and with strong evidence of non-environmental causes, through targeted gene sequencing and specific multiplex polymerase chain reaction (PCR); 3) Using multiplex ligand-dependent probe amplification (MLPA) technique to investigate the most common variants associated with DMD in Cameroon and assess their possible implication in HI in humans; 4) Performing whole exome sequencing (WES) on 2 Cameroonian multiplex families with NSHI and who tested negative for pathogenic variants in GJB2 and GJB6, to identify the underlying causative genes; 5) Performing in-depth interviews to gain an understanding of the challenges faced by people with HI in Cameroon, their understanding of the causes of hearing impairment (HI), and how challenges could be remedied to improve the quality of life of persons with HI. Results Literature reviews Our first systematic review showed that HI is a public health issue in Cameroon, especially in the elder population where the prevalence of HI is 14.8% in people aged 50 years and more. Environmental factors, including meningitis, impacted wax, and age-related disorders are the leading aetiologies of HI in Cameroon as in many other SSA countries, contributing 52.6% to 62.2% of HI cases. Hereditary HI comprises 0.8% to 14.8% of all cases in Cameroon, and in 32.6% to 37% of HI cases, the origin remains unknown. This contrasts with findings from highincome countries where hereditary HI constitutes the main aetiology of HI, contributing to approximately 50% of cases. NSHI is the most frequent clinical entity and accounts for 86.1% to 92.5% of cases of hereditary HI in the Cameroonian population. No pathogenic variant was described in GJB6 gene, and the prevalence of pathogenic variants in GJB2 ranged from 0% to 0.5%. The prevalence of pathogenic variants in other known NSHI genes was with type 2 Waardenburg syndrome, and three cases of type 2 Usher syndrome were identified in one family. By direct gene sequencing of the coding region of GJB2, no variants were found in any of the 29 families with NSHI. Additionally, through a specific multiplex PCR, the GJB6- D3S1830 deletion which contributes to 9.7% of NSHI cases in Europeans was not identified in any of the patients with HI. Subsequently, a total of 17 males with DMD from 14 families were recruited, aged 14 ± 5.1 (8–23) years. The mean age at onset of symptoms was 4.6 ± 1.5 years, and the mean age at diagnosis was 12.1 ± 5.2 years. Proximal muscle weakness was noted in all patients and calf hypertrophy in the large majority of them (88.2%; 15/17). Flexion contractures were particularly frequent on the ankle (85.7%; 12/14). Wasting of the shoulder girdle and thigh muscles was present in 50% (6/12) and 46.2% (6/13) of patients, respectively. No patient presented with HI. The MLPA found that deletions of at least one exon in DMD occurred in 45.5% of patients (5/11), while duplications were observed in 27.3% (3/11). Both variant types were clustered between exons 45 and 50, and the proportion of de novo variant was estimated at 18.2% (2/11). Whole exome sequencing We submitted DNA samples from five members of a multiplex non-consanguineous Cameroonian family segregating prelingual and progressive ARNSHI for WES. We identified novel bi-allelic compound heterozygous pathogenic variants in CLIC5 (MIM: 607293). The variants identified, i.e. the missense [NM_016929.5:c.224T>C; p.(Leu75Pro)] and the splicing (NM_016929.5:c.63+1G>A), were validated using Sanger sequencing in all seven available family members and co-segregated with HI in the three family members with HI. The three affected individuals were compound heterozygous for both variants, and all unaffected individuals were heterozygous for one of the two variants. Both variants classify as pathogenic by the American College of Medical Genetics (ACMG) guidelines for classification of variants and are absent from the genome aggregation database (gnomAD), UK10K, Greater Middle East (GME) database, and the Single Nucleotide Polymorphism Database (dbSNP), as well in 122 healthy controls from Cameroon. We also did not identify these pathogenic variants in 118 unrelated sporadic cases of NSHI from Cameroon. A second multiplex family was also screened through the use of WES, followed by direct Sanger sequencing in additional patients and control participants. We identified a heterozygous novel missense variant [NM_001174116.2:c.918G>T; p.(Gln306His)] in DMXL2 (MIM:612186) which was transmitted in an autosomal dominant manner, and co-segregates with congenital/prelingual profound to total non-syndromic sensorineural HI in a family from Cameroon. The described family showed a variable expressivity of the HI phenotype. The p.(Gln306His) variant which substitutes a highly conserved glutamine residue is predicted deleterious by various bioinformatics tools and is absent from several genome databases including genome aggregation database (gnomAD), and trans-omics for precision medicine (TOPMed) database. This variant was neither found in 121 healthy controls without personal or family history of HI, nor 112 sporadic cases of NSHI from Cameroon. Our study identified novel variants in CLIC5 and DMXL2 in two Cameroonian families, and provided only the second report of variants in these genes worldwide; thus, strengthening the case for these two genes as candidate genes for NSHI in humans. The psychosocial burden of HI We performed in-depth interviews with 10 HI professionals (healthcare workers, and educationists), and 10 persons affected by HI (persons with HI, and caregivers). The results show that in this study population, the cause of HI is attributed to a variety of causes, including genetics, environmental factors, and a spiritual curse. There were reported cases of stigma and discrimination with persons with HI in the Cameroonian population sometimes seen as having a “mental disorder”. Our participants also highlighted the difficulty that persons with HI have in accessing the necessary education and healthcare services, and suggested the need for policymakers and researchers to develop strategies to improve the social integration of persons with HI and their access to basic social services. This includes 1) Increased awareness amongst the general population, 2) the establishment of more special schools, and 3) building and equipping facilities for proper management of HI. Conclusions Our project confirms that variants in GJB2 and GJB6 genes do not contribute significantly to NSHI in the Cameroonian population. Also, variants in DMD that were shown to be associated with an increased hearing threshold in mice, do not seem to be implicated in HI in Cameroon, neither in previous human studies (although they did not objectively assess hearing using standardized testing methods). Despite the first symptoms of DMD occurring in infancy, the diagnosis is frequently made later in adolescence, indicating an underestimation of the number of cases of DMD in Cameroon. Future screening of deletions and duplications in patients from Cameroon should focus on the distal part of the DMD gene. Subsequently, this study successfully identified the candidate genes in two Cameroonian multiplex families with NSHI through the use of WES, and thus highlights the efficacy of next-generation sequencing techniques in resolving HI cases in Cameroonians and in cases where no pathogenic variants are found in common HI-genes. Additionally, our project which confirms that CLIC5 and DMXL2 genes are associated with HI in humans advocate for the inclusion of these two genes in diagnostic gene panels for NSHI in clinical settings. Last, this study shows the difficult social interaction and access to proper management faced by persons with HI in Cameroon, and highlights the need to educate populations on the causes of HI for a better acceptance of persons with HI in the Cameroonian society.

Type 2 diabetes (T2D) is currently one of the major health challenges across the globe. Lifestyle changes are a key component of T2D management and there is growing interest in low carbohydrate high fat (LCHF) diets as a potential dietary strategy to improve glycaemic control, reduce T2D medication requirements, and improve body weight and lipid profiles. However, carbohydrate restriction is controversial. Results from observational studies generally do not support the food choices associated with carbohydrate restriction while results from short-term randomised controlled trials (RCTs) are more likely to show significant benefits of LCHF diets. Additionally, both study designs have limitations and opinion on LCHF diets is polarised due to ambiguities in how to interpret the available data. Chapter 1 of this thesis reviews the impact of prospective cohort studies, randomised controlled trials, and dietary policies on current opinions towards LCHF diets for the management of T2D. Uncertainty over the safety of LCHF diets remains a concern and additional observational studies and short-term RCTs of the same quality as existing research are unlikely to add any further clarity. For this reason, research focused on understanding the underlying mechanisms of carbohydrate restricted diets may be an alternative approach to alleviate or validate some of the concerns being expressed about LCHF diets. One such mechanism is the dysregulation of glucose production via gluconeogenesis, which is a key pathology of T2D but which has been incompletely studied. Indeed, the effects of LCHF eating on gluconeogenesis in T2D patients has not yet been studied, nor has gluconeogenesis been investigated in the context of T2D remission. This is an area of interest for future research and the aim of this thesis was to conduct preliminary studies to prepare the groundwork for such studies. There is large heterogeneity in the low carbohydrate diets that have been prescribed in controlled trials and the composition and characteristics of the LCHF diets that patients are finding effective in the real world is unknown. Study 1 (Chapter 2 of this thesis) aimed to better understand the LCHF diet by investigating the diet, diabetes status, and personal experiences of T2D patients who had self-selected and followed an LCHF diet of their own accord. This study was a multi-method investigation which consisted of quantitative assessments of diet and diabetes status, as well as in-depth interviews which were analysed using qualitative methods. Results from this study will be used to inform design and protocol decisions in future controlled trial studies. Study 2 (Chapter 3 of this thesis) piloted the use of stable isotope tracers for the quantification of endogenous glucose production and gluconeogenesis in the early postabsorptive state (5 hours after a meal). For methodological reasons, prior investigations have usually measured gluconeogenesis after an overnight fast and therefore, little is known about the effects of dietary composition on gluconeogenesis within the early post-absorptive state. Study 2 quantifies gluconeogenesis 5 hours after a meal and the validity of the data is discussed. Finally, Chapter 4 outlines future perspectives for research based on findings from Chapter 2 and Chapter 3.

The Sub-Saharan African region remains the most severely affected by the HIV/AIDS epidemic. At the end of 2011, The Joint United Nations Programme on HIV/AIDS (UNAIDS) estimated that about 5% of adults were living with the HIV in this region, accounting for 69% of the global HIV prevalence. Efforts to curb the epidemic are focused on managing HIV through prevention strategies, such as advocating the use of condoms or pre-exposure or post-exposure prophylactic treatment, and prolonging life through the use of antiretroviral (ARV) therapy. Drugs used in ARV therapy target different major steps of the HIV reproductive cycle. These are nucleoside and non-nucleoside reverse transcriptase inhibitors (NRTIs/NNRTIs); fusion/entry inhibitors; integrase inhibitors; and protease inhibitors (PIs). In South Africa PIs, specifically lopinavir (LPV) boosted with another PI, ritonavir (RTV) are used in second-line ARV regimens along with a backbone of 2 NRTIs. The use of ARVs is not without issues - patients often experience side-effects to the drugs such as nausea, diarrhoea, and lipodystrophy with LPV use, which may influence their adherence to treatment and eventually lead to treatment failure. Inter-individual variability exists in patients' response to treatment despite the standard dose of 400 mg/100 mg (LPV/RTV) that is given and this may be due to differences in transport or metabolism of the drug in the liver. High plasma drug levels (associated with side-effects or toxicity) may be a result of poor metabolism or conversely, low plasma drug levels (associated with failure to suppress the virus) may be a result of extensive metabolism of the drug. Proteins involved in the disposition of LPV include the drug metabolising enzymes, CYP3A4 and CYP3A5; the hepatic uptake transporter, OATP1B1; and the efflux transporter, MRP2. Variation in the genes encoding these proteins may influence their functioning and hence LPV disposition. The aim of the study was to identify significant single nucleotide polymorphisms (SNPs) in each gene; to genotype a cohort of HIV-infected patients from Malawi and South Africa to identify the frequency of those variants; and to correlate genotypes with LPV plasma levels and other clinical parameters.

Includes bibliographical references. / The history of the nature conservation in South Africa’s protected areas is marked by the unfortunate reality of forced removals and land dispossession. Ultimately landlessness created an unequal society in terms of land holding, use and ownership. Nature conservation was also not spared. The land reform program in South Africa more especially the land restitution in the Kruger National Park re-defined the relationship between nature and society. However, such redefinition of nature and society takes place at the confluence of neo-liberalisation of nature and neoliberal land reform. The thesis provides insight into variants of neo-liberalism which point to ways in which nature conservation is increasingly being incorporated into market conditions and ideals. In the case of South Africa, the neo-liberalisation of nature takes place through a market-based approach to land reform. This brings together two threads of neo-liberalism, namely, the neo-liberalisation of nature and neo-liberal land reform. The study focuses on the land restoration debate which revolves around whether it is feasible to restore all land that was lost as a result of apartheid’s discriminatory practices; the appropriate method for achieving an equitable land restitution; and the method by which such restitution can be achieved with due consideration to all other national imperatives and long term goals.