Computer-based analysis of genotype-phenotype correlation in selected genes associated with epilepsy

In this thesis, the potential biological impacts of yanlış anlamlı single-nucleotide polymorphisms (nsSNPs) classified as variants of uncertain significance (VUS) within the CHRNA2, CHRNA4, and CHRNB2 genes, which are associated with Autosomal Dominant Nocturnal Frontal Lobe Epilepsy (ADNFLE), were comprehensively evaluated through a multilayered and systematic In silico analysis strategy. These genes encode the α2, α4, and β2 subunits of the neuronal nicotinic acetylcholine receptor (nAChR), which plays a crucial role in modulating cholinergic transmission within the central nervous system. Therefore, investigating the structural and functional consequences of variations in these genes is critical for understanding the molecular mechanisms underlying epileptogenesis. In the initial phase, variants classified as "yanlış anlamlı" and "VUS" were extracted from the ClinVar database. Among the identified variants—357 for CHRNA4, 288 for CHRNA2, and 275 for CHRNB2—those with REVEL scores ≥0.9 and categorized as "pathogenic (P)" by AlphaMissense were selected for further analyses (CHRNA4: 38; CHRNA2: 51; CHRNB2: 37 variants). The selected variants were evaluated using PolyPhen-2 (HDIV/HVAR), SIFT, ConSurf, gnomAD, SpliceAI, DynaMut2 and HOPE tools. Each variant was comprehensively assessed in terms of pathogenicity tendency, evolutionary conservation, allele frequency, structural integrity, stability alterations, potential splice-site effects, and interactions with ligand-binding regions. The findings revealed consistent patterns across all three genes. Nearly all high-confidence variants were classified as "damaging" by PolyPhen-2 and "deleterious" by SIFT, while ConSurf analysis showed that the majority were located in highly conserved positions. gnomAD screening indicated that these variants occurred at rare or ultra-rare frequencies (≈10⁻⁷–10⁻⁵), and a subset had not been previously reported. DynaMut2 analyses demonstrated predominantly negative ΔΔG values, suggesting destabilizing effects on protein structure. HOPE modeling revealed frequent biophysical mechanisms, including changes in charge, volume, and hydrophobicity, as well as proline/glycine-induced backbone constraints and disruption of disulfide bridges. Although SpliceAI scores were generally low, measurable acceptor/donor gain–loss probabilities were detected at certain exonic positions. Structural mapping indicated that high-risk variants were enriched within the extracellular ligand-binding interface and TM3–TM4/TM4 regions, with some clustering at multimer contact and transmembrane core sites. Integration of signals from multiple bioinformatic tools (≥6/8 concordance) enabled the prioritization of a candidate variant panel for future functional validation and potential clinical reclassification. In conclusion, this thesis provides a comprehensive framework combining structural, functional, and population-based analyses of VUS-type nsSNPs within CHRNA2, CHRNA4, and CHRNB2 genes. The findings contribute to elucidating the molecular basis of rare epilepsy forms, support the integration of In silico evidence into clinical variant interpretation, and highlight the power of multilayered computational approaches in the molecular characterization of neurogenetic disorders.