20.02.2024Utilizing Genetic Linkage and Rare Variant Association Studies to Unravel Novel Disease Etiologies

This thesis aims to use bioinformatic approaches to elucidate the underlying genetic causes of three different diseases. Our studies are mostly based on novel sequencing technologies (Chapter 1) and samples from patients diagnosed at the Institute of Medical Genetics and the Moyamoya Center at the Pediatric University Hospital in Zürich.

In Chapter 2, we study Moyamoya angiopathy (MMA), a condition characterized by stenosis of the aortal carotid of the patients and occlusion of cerebral arteries, the main cause of ischemic and hemorrhagic stroke. It has been mainly studied in Asian adult populations, where a founder variant was found in the RNF213 gene that is absent in European populations.
Our methods consisted of a case-control study in pediatric Moyamoya Angiopathy patients of mostly European ancestry. The whole exome was sequenced from blood samples in a pediatric cohort of 105 MMA patients and 234 healthy control individuals (88% Europeans). After quality control and excluding common variants (MAF >2%) and homologous regions, we performed a rare variant enrichment test using the R package SKAT. Without an independent patient cohort, we verified our results by repeating the test using the same Moyamoya cohort against a different control cohort of 307 individuals and again with individuals predicted to be exclusively of European origin. Our results revealed significant enrichment of rare variants in the known MMA gene RNF213, ranked among the most enriched in the three tests, and another 28 significantly enriched genes (p-value <0.01), four of them already associated with MMA (RNF213, NF1, OBSCN). Together with RNF213, LAMA5 is the only common gene in the three tests, standing as a strong novel candidate gene for MMA susceptibility. The most significantly enriched in our largest cohort test was COL18A1, which was previously associated with MMA exclusively in Trisomy 21 patients.

Chapter 3 focuses on the kyphoscoliosis disease, specifically in an early-onset, fully penetrant, monogenic autosomal dominant familial type, for which the pathogenic variant is unknown. Kyphoscoliosis is a spine deformity characterized by its curvature in the sagittal and coronal planes in more than 10 degrees measured by the Cobb angle. It is caused by the abnormal segmentation of the spine during embryonic development. Previous familial studies have shown many variations of the disease between families, differing in age at onset, severity, incidence, and inheritance model. The
inheritance can be autosomal dominant, X-linked, multigenic, or multifactorial. Despite the previous efforts, no major susceptibility variants have been suggested.
We performed genome-wide linkage analysis using single nucleotide polymorphism (SNP) microarrays, which detected a specific 30 cM linkage region in chromosome 2 with a maximum likelihood (LOD score) of 2.4. Next, we sequenced the exome (WES) of 11 individuals in the family, with six affected and five unaffected. By filtering uninformative variants and finding shared haplotypes among affected individuals, we confirmed the linked region identified by the previous linkage analysis and narrowed it down to six genes.

Finally, in Chapter 4, we started a study on hereditary breast cancer. Deleterious variants in the tumor suppressors BRCA1 and BRCA2 predispose about 70% lifetime risk of developing breast and other cancers. So far, 19 genes have been robustly associated with hereditary breast and ovarian cancer (HBOC) susceptibility. However, pathogenic variants in HBOC genes are present in only 5-10% of diagnosed cases.

In this project, we are using whole genome sequencing data and applying bioinformatic tools to find novel HBOC genes to better diagnose patients and understand the disease.
Our cohort consisted of 52 hereditary breast cancer patients with at least one further family member affected by cancer. All patients were previously screened for the 19 known HBOC pathogenic variants, with only one pathogenic variant found in a gene that was not investigated in the diagnostic workup. The control cohort consisted of 148 healthy controls with no family history of cancer.
In our first analysis, we performed a rare variant enrichment testing similar to the Moyamoya study in Chapter 2, focusing on exonic regions, ±20 bp. We performed statistical collapsing tests using the R package SKAT in eight sets of tested genes and variants. Our preliminary results revealed five interesting significantly enriched gene.

DOI: 10.5167/uzh-257321, Dissertation, 2024, University of Zurich, Mathematisch-naturwissenschaftliche Fakultät.

Elena M. Cabello Ferrete