Exploring the Genetic Landscape of Psychotic Disorders Using Trio-based Short and Long Reads Sequencing

Psychiatric disorders such as schizophrenia and bipolar disorder have a complex genetic basis, shaped by both common and rare variation. Short-read whole genome sequencing (SR-WGS) has remained the standard for variant discovery, although it has known limitations in resolving structural variation and repetitive regions. While large-scale genome-wide association studies (GWAS) have identified many common risk loci, trio-based analyses complement these approaches by enabling the detection of rare inherited and de novo variants that are not readily captured by association studies alone. In this thesis, genomic variation was analysed in a psychotic cohort from the MNESYS Spoke 5 project using a combination of short-read and Oxford Nanopore long-read sequencing. Short- read data were used as the primary approach for identifying single nucleotide variants (SNVs) and small indels, leading to the prioritisation of several candidate variants, including de novo and compound heterozygous events. These findings reinforce the strength of short-read sequencing for high-confidence detection of small variants and its continued relevance in clinical and research settings. Long-read sequencing was applied to a subset of samples, particularly those that remained unresolved after short-read analysis, with a focus on structural variant (SV) and copy number variant (CNV) detection. A multi-step workflow incorporating dual-caller consensus, trio-aware filtering, and re-alignment to the T2T-CHM13 reference was used to reduce false positives and refine candidate events. This approach resulted in a marked reduction in technological artefact calls and enabled the identification of large copy number variants. A comparison between sequencing technologies showed substantial agreement in coding regions, alongside an increased number of variants detected by long reads in non-coding and structurally complex regions. In selected cases, in silico functional prediction (AlphaGenome) was used to provide additional context, although these results were interpreted cautiously. Overall, the results highlight the heterogeneous genetic architecture of neuropsychiatric disorders and support the use of combined sequencing strategies. While short-read sequencing remains acceptable for SNV detection, long-read approaches provide complementary value, particularly for structural variation and challenging genomic regions. This work outlines apractical framework for integrating both technologies in cohort-scale studies while also highlighting current limitations in variant interpretation and clinical translation.