Résumé
In a context where whole-genome analysis is becoming increasingly accessible, the facilitated discovery of intronic variants raises questions about their interpretation. New predictive bioinformatics tools are regularly proposed to help classify them. In this project, we compared the predicted and observed impact on patient RNA in vivo of nine intronic variants identified through whole-genome analysis (LBMMS AURAGEN) or mini-exome analysis (CHU Montpellier) in patients with neurodegenerative diseases, involving various genes related to these pathologies. We collected blood samples in PAX gene tubes (for RNA stabilization) to determine the splicing profiles in leukocytes of the genes (all ubiquitously ex- pressed) carrying the different variants, using RT-PCR and Sanger sequencing. This approach allowed us to compare in vivo observations with algorithm predictions. A functional approach based on the use of a reporter minigene in HeLa cells, also allowed us to explore the mechanisms involved for one of the variants affecting a cis-regulatory splicing sequence by modulating the expression of different splicing factors (overexpression/depletion). The molecular and clinical diagnoses established by these functional analyses are of great importance, as they guide management decisions, notably enabling the early implementation of genetic counseling, which can sometimes lead to prenatal or pre- symptomatic diagnosis for relatives. Currently, variants considered pathogenic due to splicing alterations (according to ACMG classification) are rarely located outside canonical splice sites (+/- 1 or 2) due to the requirement for transcript analysis to confirm their deleterious effect. It is crucial to refine our understanding of the various available predictive programs and to master the design and implementation of experimental procedures to implicate different types of intronic mutations in patient pathology.