Mendelian randomisation and complementary informatics analyses of environmental and molecular determinants in complex respiratory and neurological disorders using large-scale biobank data

Mendelian randomisation (MR) was combined with orthogonal validation to identify the
molecular drivers of complex diseases, focusing on pulmonary arterial hypertension (PAH) and amyotrophic lateral sclerosis (ALS). Large-scale biobank resources, including UK Biobank, FinnGen, and deCODE were used to develop a reproducible framework that integrates an initial hypothesis-free MR screen with orthogonal analyses, including rare genetic variant association testing, transcriptomic analyses, clinical measurements, and functional validation. Brief summary of each chapter: A study of healthcare workers with nonhospitalised COVID-19 revealed two symptom patterns of long and short duration; distance covered by walking or running was the only physical activity measure associated with persistent symptoms. In PAH, a hypothesis-free MR screen of 575 circulating metabolites identified five candidates whose serum concentration was causally linked to disease risk; clinical validation confirmed serine as protective and homostachydrine as harmful. These conclusions were supported by a rare variant analysis implicating serine synthesis regulation in disease risk. Study of ALS analyses extended this framework across three dimensions. First, sex-stratified MR and UK Biobank follow-up demonstrated that extreme physical activity increases ALS risk in males, particularly those with genetically inhibited mTOR signalling, implicating neuromuscular junction dysfunction in susceptible individuals. Second, metabolite analyses highlighted the protective effect of carnitines and the disease risk associated with DSGEGDFXAEGGGVR, a peptide linked to the neuronal glucose transporter GLUT3; overall, impaired metabolic flexibility in motor neurons was found to increase ALS risk. Third, a neurofilament-guided screen of circulating proteins prioritised plasma BRK1 as protective against ALS risk, alongside CD14, OMG, SMOC1, DPY30, and DTNB, implicating cytoskeletal regulation, immune activation, glial biology, and
neuromuscular junction integrity in ALS pathogenesis. In conclusion, combining MR with
orthogonal validation enhances causal inference, reveals sex-specific and heterogeneous
mechanisms in PAH and ALS, and establishes a framework with translational potential for biomarker discovery and therapeutic targeting.