Universität zu Köln

Brünger, Tobias ORCID: 0000-0003-2591-432X (2023). Bioinformatic approaches to determine pathogenicity and function of clinical genetic variants across ion channels and neurodevelopmental disorder associated genes. PhD thesis, Universität zu Köln.

PDF Dissertation_Tobias_Bruenger_KUPS_signed.pdf - Accepted Version Download (37MB)

Abstract

Clinical genetic testing for rare monogenic diseases has the scope of identifying the disease-causing variants. Identification of the molecular etiology of the disease can already today improve clinical care and is essential for the administration of precision medicines that are currently in development for many disorders. However, distinguishing pathogenic variants from benign genetic variants remains a challenge – in particular for missense variants where a single amino acid is substituted. The effects of a pathogenic variant on the protein function, for example, whether it causes a gain (GoF) or a loss (LoF) of the protein function, is most of the time not understood since most genetic variants are ultra-rare and have not been molecularly tested. In particular, for genes associated with severe developmental disorders, first-generation symptomatic treatments offer often only limited relief. Consequently, the development and application of targeted treatments that promise improvement is urgently needed. Identifying the disease-causing pathogenic and predicting their function is crucial as targeted therapies can only be administered to patients with classified pathogenic variants whose functional effects are known to avoid adverse treatment outcomes. In this dissertation, I present bioinformatic approaches to enhance the assessment of variant pathogenicity and understanding of the functional effects of genetic variants. The developed approaches were applied on an exome-wide scale using public datasets and for selected disorders for which I had expert-curated clinical-genetic data available from collaborators. The major focus of this thesis is on genes implicated in neurodevelopmental disorders and diseases associated with ion channel dysfunction for which collaboration with other research groups enabled the aggregation of required genetic, clinical, and functional datasets to develop and test the bioinformatic approaches. In the first study (Bruenger and Ivaniuk et al., in preparation for submission to Genetics in Medicine), we developed a novel approach to extend the application of current variant interpretation guidelines as proposed by the American College of Medical Genetics and Genomics (ACMG). Currently, a major limitation of interpreting the pathogenicity of variants with the ACMG guidelines presents the rare applicability of some of the proposed evidence criteria. We evaluated the potential of incorporating individual pathogenic variants observed in paralogous genes to extend the applicability of two criteria of the guidelines. Our results demonstrated that pathogenic variants in evolutionarily conserved paralogous genes can serve as evidence for a variant's pathogenicity and thus extend the current criteria's applicability by more than four times. We further explored whether the selection of the paralogous pathogenic variants can be improved by incorporating phenotype information. We assembled a clinically well-defined cohort of patients with variants in voltage-gated sodium channels (VGSC) and identified phenotype correlations among paralogous genes based on the shared variant properties. By integrating these phenotype correlations into our proposed extension of the ACMG criteria, we demonstrated an enhanced ability to provide evidence for the pathogenicity of genetic variants in VGSC-encoding genes. In the second study (Brunklaus, Feng, and Bruenger et al., Brain, 2022), we examined whether experimentally obtained functional effects of variants in one VGSC encoding gene could predict function in conserved variants in paralogous genes with high sequence similarity. We aggregated 437 in-vitro functionally tested variants from an intensive literature search and found that the functional effect across conserved variants in paralogous genes was conserved in 94% of cases. Our findings represent the first GoF versus LoF topological map of VGSC proteins, which could guide precision therapy as functionally tested variants are rare across VGSC. We integrated our findings into a publicly accessible webtool (http://SCN-viewer.broadinstitute.org) to facilitate functional variant interpretation across VGSC. In the third study (Bruenger et al., Brain, 2022), we systematically identified biological properties associated with variant pathogenicity across all major voltage and ligand-gated ion-channel families. We discovered and independently replicated that several pore residue properties and proximity to the pore axis were significantly enriched for pathogenic variants compared to population variants across all ion channels. Using a newly developed structural framework, we provide quantitative evidence that variants at the pore showed the strongest pathogenic variant enrichment. Moreover, we found that a hydrophobic pore environment was most strongly associated with variant pathogenicity. Finally, we showed that the identified biological properties correlated with in-vitro functional readouts from 679 variants and clinical phenotypes in 1,422 patients with neurodevelopmental disorders which were collected through collaboration with other research groups. In summary, we identified biological properties associated with ion-channel malfunction and show that these are correlated with in vitro functional readouts and clinical phenotypes in patients with neurodevelopmental disorders. Our results suggest that clinical decision support algorithms that predict variant pathogenicity and function are feasible in the future. In the fourth study (Iqbal and Bruenger et al., Brain, 2022), we developed a novel consensus approach that combines evolutionary and population-based genomic scores to identify 3D essential sites (Essential3D) on protein structures encoded by genes associated with neurodevelopmental disorders (NDDs). NDDs encompass severe clinical conditions caused by pathogenic variants in different genes. However, many of those genes were just recently associated with NDDs and are not well studied. We identified 14,377 Essential3D sites on protein structures encoded by 189 genes and found that these sites were eight-fold enriched for pathogenic versus population controls in an independent cohort of over 360,000 patient and population variants. The Essential3D sites offer insights into molecular mechanisms of protein function, such as key protein-protein interaction sites. The provided annotations are available at https://es-ndd.broadinstitute.org and will guide clinical variant interpretation. In summary, within these major studies in my Ph.D., we aggregated genetic, clinical, and functional datasets and developed bioinformatic approaches to enhance the assessment of variant pathogenicity and improve understanding of the functional effects of genetic variants on protein function. The advances made during my Ph.D. research demonstrate the power of integrating multiple data sources to study novel genetic variants and their implication for rare monogenic diseases. Our approaches specifically improve variant function and pathogenicity assessment in genes implicated in several severe diseases for which currently applied first-generation therapies cannot adequately lower the disease burden. Thus, our results contribute to a new era in precision medicine, where personalized treatments and improved clinical care become increasingly accessible to patients. Finally, the annotations developed in these can serve as a foundation for further studies, including the application of machine learning methods to predict variant pathogenicity and protein functional effects more accurately.

Item Type:	Thesis (PhD thesis)
Creators:	Creators Email ORCID ORCID Put Code Brünger, Tobias tbrueng1@uni-koeln.de orcid.org/0000-0003-2591-432X UNSPECIFIED
URN:	urn:nbn:de:hbz:38-714266
Date:	25 September 2023
Language:	English
Faculty:	Faculty of Mathematics and Natural Sciences
Divisions:	Cologne Center for Genomics
Subjects:	Life sciences
Uncontrolled Keywords:	Keywords Language Bioinformatics English Genetics English Variant classification English
Date of oral exam:	25 September 2023
Referee:	Name Academic Title Nothnagel, Michael Prof. Dr. Hofmann, Kay Prof. Dr. Krawitz, Peter Prof. Dr.
Refereed:	Yes
URI:	http://kups.ub.uni-koeln.de/id/eprint/71426