Understanding population-level genetic interactions for neurodegeneration in the Indian Population

People worldwide suffer from a plethora of neurodegenerative diseases, yet the pathogenesis of most of these diseases is not fully known. Neurodegeneration is the result of the loss of function of neurons and their progressive atrophy. Many of the body’s activities, such as movement, balance, memory, and cognition, can get impaired due to neurodegeneration, because of which patients can become entirely dependent on caregivers. There is the Global Burden of Disease Study (GBD), where over 6500 researchers from more than 155 countries conduct a collaborative effort to quantify the burden of various diseases. As per a GBD study, as of 2016, Alzheimer’s disease and Dementia contribute 10.4% to disability-adjusted life years (DALYs). This metric refers to the sum of years lived with disability and years lost due to premature death. Neurodegeneration is primarily heritable and is a multigenic, multifactorial, and complex trait. One of the most significant risk factors is increasing age. Because neurodegenerative diseases manifest in mid to late life, as the population ages, the number of cases is expected to rise unless prophylactic and therapeutic interventions are discovered. While past studies have indicated that maintaining a healthy lifestyle impedes disease progression, individuals already carry a certain risk of predisposition to neurodegeneration depending on their genetic makeup. It is known that the human apolipoproteinE (APOE) gene, along with other genetic risk loci, confer high risk for neurodegeneration. But they account for only a certain percentage of heritability (for instance, 0.24-0.53% in case of Alzheimer’s disease). Thus, it is essential to identify additional genetic risk factors, their interactions among themselves as well as the environment, and the pathways involved to explain more of this heritable condition. While evidence suggests APOE gene to be the major predisposing factor for neurodegeneration, the mechanism by which it contributes to onset and progression is unknown, and that is why no medication has been discovered that can prevent inception or stop disease advancement. Thus, an important research question is to identify the genetic loci working in conjunction with APOE in humans via their functions in putative pathways which have been hypothesized, on the basis of mouse-model as well as functional studies, to play a role in manifesting cognitive impairment in neurodegeneration. Large-scale human population-based association studies with samples that are indeed representative of the population can provide more reliable evidence of disease pathogenesis in a cost-effective manner. My work focusses on the Indian and European population with the critical objectives of implicating genetic variants associated with APOE in symptoms of neurodegeneration and determining their downstream causal pathways considering underlying interactions comparing results across global cohorts to identify population-specific differences in genome architecture accounting for differences in disease manifestation. 

Most population-based association studies have focussed on cohorts with western ancestry, but no human population-based study has been focussed on the Indian population. Due to evolutionary and environmental factors, the genetic makeup of the Indian population is expected to be different. So far, no such large-scale studies based entirely on the Indian population have been conducted. This study can set an example and identify novel gene locations and susceptibility loci for neurodegeneration specific to the Indian population compared to European population. Uncovering such novel locations can unleash preventive and therapeutic strategies that could be better suited in terms of efficacy for the Indian population and pave the way to precision medicine. Interactions at the population scale to understand the disease etiology beyond the contribution from single genetic-variant effects are rarely studied. While elucidating the disease mechanism, this interaction network knowledge at the genetic level will also help to uncover additional genetic risks that can explain the heritability of neurodegeneration better and the reasons for differential results (if any) among populations. 

As part of the GenomeIndia Consortium initiative, I contributed to one of the largest whole-genome sequencing efforts in India, involving over 9,000 healthy individuals across diverse Indian populations. My work specifically focused on leading the construction and validation of the phased GenomeIndia (GI) haplotype reference panel and providing downstream LD resources. 

We carried out linkage disequilibrium (LD) characterization, imputation benchmarking, and shared-site overlap analyses to evaluate the performance and population specificity of the panel. By leveraging high-coverage whole-genome sequencing data, we helped develop a population-specific genotype imputation resource that captures the unique haplotype architecture and genetic diversity of Indian populations. Our analyses demonstrated that the GenomeIndia reference panel substantially improves imputation accuracy compared to existing global reference panels, leading to increased power for downstream analyses such as genome-wide association studies, improved fine-mapping resolution, and better transferability of polygenic risk scores in South Asian populations. This work contributes toward addressing the longstanding Eurocentric bias in public genomic datasets and enables more accurate and equitable genomic research for Indian and South Asian ancestries.

Through this first-of-its-kind study, we identified both known and novel variants influencing cognition and metabolic traits in rural Indian populations, while providing the first causal evidence linking lipid metabolism to domain-specific cognitive outcomes. We collected genomic and phenotypic data from over 5000 individuals residing in rural South India. We developed a new population-specific ancestry-matched reference haplotype-panel for genomic imputation specific to improve the accuracy of genetic analysis in this health disparate underrepresented population.  Using genome-wide analyses, we identified genetic variants associated with cognitive domains (like memory and attention) and metabolic traits (like cholesterol and insulin resistance). Our study uncovered one novel memory-associated locus (FLRT2-LINC02328) and 17 novel cardiometabolic loci, alongside population-specific haplotype structures in genes like AMIGO1 and ZPR1. We implemented statistical methods, including mendelian randomization, to test whether metabolic factors can cause changes in cognitive function. Our results showed that high triglycerides, elevated visceral adiposity, and insulin resistance exerts an adverse causal influence on cognitive domains such as memory, attention, and visuospatial abilities, highlighting the APOC3-APOA4-APOA5-ZPR1-BUD13 gene cluster as a key candidate for understanding this systemic metabolic-cognitive nexus.

We identify genetic variants acting in conjunction with APOE  to influence cognition while teasing out effects of other age-associated metabolic risk factors like dyslipidemia and glycemic imbalance co-occuring with age based on a whole exome-wide study on 157,160 individuals. We detect 18 previously unknown genetic variants in this context, that could influence five cognition domains. We also identify overlapping association hits when we restrict our study to coding regions only. We further identify two lipid-transport genes - APOC1 and LRP1 associated with two domains of cognition, namely complex processing speed and visual attention. We also find hitherto unknown interactions between APOE and other variants in AMIGO1 and other loci, which could affect episodic memory, simple processing speed, and visual attention, even while controlling for metabolic risks. Besides these independent effects, we uncover mediating and pleiotropic effects of these variants on cognition and metabolic risks. All the significant variants and genes we uncover show high expression in the brain and are essential for functions that correlate with neuronal mechanisms underpinning cognition. From the distinct association results for each of the cognitive domains, our results emphasize the need to study the genetic architecture of cognitive functioning at a more granular domain-specific level.

Our preliminary work on epistasis analysis was carried out for genome-wide body mass index (BMI) associated SNPs in Alzheimer’s Disease Neuroimaging Initiative (ADNI), with top significant interacting SNPs followed up in the UK Biobank imputed genotype dataset with more than 188,000 samples. We tested a consensus of nine different epistatic tools which implemented exhaustive, heuristics, stochastic  and machine-learning approaches, and discovered two pairwise epistatic interactions, one between rs2177596 (RHBDD1) and rs17759796 (MAPK1), and the other between rs1121980 (FTO) and rs6567160 (MC4R). These two pairwise epistatic interactions signify neuronal influence in obesity. This study suggested concerted expression of associated genes in metabolic abnormalities characterized by obesity, and highlighted the importance of understanding the effect of concurrently interacting genetic loci in disease aetiology, beyond single locus effects.