Cystic brosis (CF) is a life-threatening but rare genetic disorder, prevalent among individuals all over the world, and caused by mutations in the CFTR gene. Mostly, CF screening programs involve immunoreactive trypsinogen (IRT) testing and targeted CFTR variant panels. However, these methods produce a signicant number of false positives and leave a subgroup of IRT-positive but CF-negative infants. The genetic background of these individuals remains poorly understood which raises important questions about alternative predisposition pathways, particularly those linked to pancreatic disease. This thesis focuses on whole-exome sequencing data analysis of a pediatric cohort of 103 IRT-positive non-CF infants (false positives) and 119 IRT-negative controls (true negatives/very negatives). After ltering at variant-level and quality control measures, a total of 1125 variants were retained for statistical analysis. Fisher's exact test prioritized multiple variants with nominal associations, and false discovery rate (FDR) correction was used to control for multiple testing. At a stringent threshold of p < 0.001, forty-nine variants remained signicant which were dominated by PRSS2 (22 variants) and PRSS1 (19 variants), with additional signals in KRT18 (7 variants) and CFTR (1 variant). Notably, KRT18 signals few variants but three of them achieved extreme signicance (p < 1 × 10−11). This suggests that epithelial cytoskeletal pathways play an important role as modiers in pancreatic physiology. To complement single-variant analyses, gene-based burdens including Variable Threshold (VT), Sequence Kernel Association Test (SKAT), SKAT-Optimized (SKAT- O) and Combined Multivariate & Collapsing (CMC) tests were applied. These methods (SKAT and SKAT-O) consistently suggest that PRSS1 and PRSS2 show the strongest gene-level associations. These results were tested for Bonferroni correction which conrmed their role as robust modier loci. Additionally, the VT method suggested CLDN2 and MORC4 as strong signals, while CMC highlighted CCL2 and TRPV6 as strong candidates. Together, these results highlighted the central role of modier genes while nominating additional loci for future validation. Functional annotation using ClinVar, CFTR2, and gnomAD databases revealed that most signicant associations involved rare alleles (AF < 0.01), which supported their likely pathogenic and disease-modier role. However, mostly variants which were already known as disease-causing or likely disease-causing appeared in the CFTR gene, but most of them were not statistically common in cases. Additionally, variants with moderate eects appeared in the PRSS1 and PRSS2 genes. iii Together, these ndings suggest that false-positive infants harbor genetic variants in pancreatic enzyme genes and epithelial structural genes that alter pancreatic function and related disorders. This work also suggests that mutated CFTR is a major candidate gene of CF, but other candidate genes such as KRT18, PRSS2, PRSS1, CLDN2, MORC4, CCL2, and TRPV6 are also important to study closely. This thesis provides novel insights into the genetic architecture of CF-related heterogeneity. Ultimately, the thesis contributes to refining early screening accuracy, reducing diagnostic uncertainty, and broadening our understanding of CF-related genetic architecture.
La brosi cistica (CF) è una malattia genetica rara ma potenzialmente letale, diusa in tutto il mondo e causata da mutazioni nel gene CFTR. Nella maggior parte dei casi, i programmi di screening per la CF prevedono il test della tripsina immunoreattiva (IRT) e pannelli mirati di varianti CFTR. Tuttavia, questi metodi producono un numero signicativo di falsi positivi e risultano in un sottogruppo di neonati IRT-positivi ma CF-negativi. Il prolo sfondo genetico di questi individui rimane in gran parte inesplorata, ponendo l'attenzione su potenziali meccanismi di predisposizione alternativi, con un focus particolare sulle patologie pancreatiche. Questa tesi si concentra sull'analisi dei dati di sequenziamento dell'esoma completo (WES) in una coorte pediatrica composta da 103 neonati IRT-positivi non CF (falsi positivi) e 119 controlli IRT-negativi (veri negativi). Dopo il ltraggio a livello di varianti e l'applicazione di misure di controllo di qualità, 1125 varianti sono state incluse nelle analisi. Il test esatto di Fisher ha individuato alcune associazioni nominali, corrette per test multipli mediante FDR. A una soglia stringente di p < 0.001, quarantanove varianti sono risultate signicative: dominate da PRSS2 (22 varianti) e PRSS1 (19 varianti), con ulteriori segnali in KRT18 (7 varianti) e CFTR (1 variante). In particolare, i segnali di KRT18 hanno mostrato poche varianti ma tre di esse hanno raggiunto un'estrema signicatività (p < 1×10−11). Questo suggerisce che le vie cito- scheletriche epiteliali svolgono un ruolo importante come modicatori nella siologia pancreatica. Per completare le analisi a livello di singola variante, sono stati applicati test di carico genico, tra cui Variable Threshold (VT), Sequence Kernel Association Test (SKAT), SKAT-Optimized (SKAT-O) e Combined Multivariate and Collapsing (CMC). Questi metodi (SKAT e SKAT-O) hanno evidenziato che PRSS1 e PRSS2 mostrano le associazioni geniche più forti. Tali risultati hanno superato la correzione di Bonferroni, confermando il loro ruolo come loci modicatori robusti. Inoltre, il metodo VT ha suggerito CLDN2 e MORC4 come segnali forti, mentre il CMC ha evidenziato CCL2 e TRPV6 come candidati rilevanti. Insieme, questi risultati sottolineano il ruolo centrale dei geni modicatori e nominano ulteriori loci per future validazioni. L'annotazione funzionale mediante i database ClinVar, CFTR2 e gnomAD ha rivelato che la maggior parte delle associazioni signicative riguardava alleli rari (AF < 0.01), a supporto del loro probabile ruolo patogeno o di modicatori di malattia. Tuttavia, la maggior parte delle varianti già note come patogenetiche o probabilmente patogenetiche si è localizzata nel gene CFTR, ma poche di esse sono i risultate statisticamente comuni nei casi. Inoltre, varianti con eetti moderati sono state osservate nei geni PRSS1 e PRSS2. Nel complesso, questi risultati suggeriscono che i neonati falsi positivi portano varianti genetiche in geni degli enzimi pancreatici e in geni strutturali epiteliali che alterano la funzione pancreatica e patologie correlate. Questo lavoro conferma CFTR come gene principale della CF, ma indica anche che altri geni candidati come KRT18, PRSS2, PRSS1, CLDN2, MORC4, CCL2 e TRPV6 meritano studi approfonditi. La tesi fornisce nuove prospettive sull'architettura genetica dell'eterogeneità correlata alla CF e contribuisce a migliorare l'accuratezza dello screening precoce, ridurre l'incertezza diagnostica e ampliare la comprensione dei meccanismi genetici della malattia.
Analisi mediante sequenziamento dell’intero esoma nei neonati con risultato falso positivo allo screening per la fibrosi cistica.
SARFRAZ, UMMARA
2025/2026
Abstract
Cystic brosis (CF) is a life-threatening but rare genetic disorder, prevalent among individuals all over the world, and caused by mutations in the CFTR gene. Mostly, CF screening programs involve immunoreactive trypsinogen (IRT) testing and targeted CFTR variant panels. However, these methods produce a signicant number of false positives and leave a subgroup of IRT-positive but CF-negative infants. The genetic background of these individuals remains poorly understood which raises important questions about alternative predisposition pathways, particularly those linked to pancreatic disease. This thesis focuses on whole-exome sequencing data analysis of a pediatric cohort of 103 IRT-positive non-CF infants (false positives) and 119 IRT-negative controls (true negatives/very negatives). After ltering at variant-level and quality control measures, a total of 1125 variants were retained for statistical analysis. Fisher's exact test prioritized multiple variants with nominal associations, and false discovery rate (FDR) correction was used to control for multiple testing. At a stringent threshold of p < 0.001, forty-nine variants remained signicant which were dominated by PRSS2 (22 variants) and PRSS1 (19 variants), with additional signals in KRT18 (7 variants) and CFTR (1 variant). Notably, KRT18 signals few variants but three of them achieved extreme signicance (p < 1 × 10−11). This suggests that epithelial cytoskeletal pathways play an important role as modiers in pancreatic physiology. To complement single-variant analyses, gene-based burdens including Variable Threshold (VT), Sequence Kernel Association Test (SKAT), SKAT-Optimized (SKAT- O) and Combined Multivariate & Collapsing (CMC) tests were applied. These methods (SKAT and SKAT-O) consistently suggest that PRSS1 and PRSS2 show the strongest gene-level associations. These results were tested for Bonferroni correction which conrmed their role as robust modier loci. Additionally, the VT method suggested CLDN2 and MORC4 as strong signals, while CMC highlighted CCL2 and TRPV6 as strong candidates. Together, these results highlighted the central role of modier genes while nominating additional loci for future validation. Functional annotation using ClinVar, CFTR2, and gnomAD databases revealed that most signicant associations involved rare alleles (AF < 0.01), which supported their likely pathogenic and disease-modier role. However, mostly variants which were already known as disease-causing or likely disease-causing appeared in the CFTR gene, but most of them were not statistically common in cases. Additionally, variants with moderate eects appeared in the PRSS1 and PRSS2 genes. iii Together, these ndings suggest that false-positive infants harbor genetic variants in pancreatic enzyme genes and epithelial structural genes that alter pancreatic function and related disorders. This work also suggests that mutated CFTR is a major candidate gene of CF, but other candidate genes such as KRT18, PRSS2, PRSS1, CLDN2, MORC4, CCL2, and TRPV6 are also important to study closely. This thesis provides novel insights into the genetic architecture of CF-related heterogeneity. Ultimately, the thesis contributes to refining early screening accuracy, reducing diagnostic uncertainty, and broadening our understanding of CF-related genetic architecture.| File | Dimensione | Formato | |
|---|---|---|---|
|
517220-UmmaraSarfraz_pdfa.pdf
non disponibili
Descrizione: Th master's thesis investigated genetic factors associated with false-positive newborn screening for cystic fibrosis using whole-exome sequencing (WES) data from 222 infants. I performed bioinformatics analysis, variant annotation, and statistical testing
Dimensione
6.38 MB
Formato
Adobe PDF
|
6.38 MB | Adobe PDF | Richiedi una copia |
È consentito all'utente scaricare e condividere i documenti disponibili a testo pieno in UNITESI UNIPV nel rispetto della licenza Creative Commons del tipo CC BY NC ND.
Per maggiori informazioni e per verifiche sull'eventuale disponibilità del file scrivere a: [email protected].
https://hdl.handle.net/20.500.14239/37325