New technologies allow the exploration of the whole genome to identify genetic variants associated with various phenotypes, in particular diseases. Bioinformatics aims at helping to answer these questions. In the context of my PhD thesis, I have first developed a new software allowing to measure with a good precision the number of really independent genetic markers present in a set of markers genotyped in a given population. This algorithm relies on the Shannon's entropy contained within these markers and on the levels of mutual information computed from the pairs of SNPs chosen in a given window of consecutive SNPs, the window size is a parameter of the program. I have shown that the number of really independent markers become stable as soon as the population is homogeneous and large enough (N > 60) and as soon as the window size is large enough (size > 100 SNPs). This computation may have several applications, in particular the diminution of the Bonferroni threshold by a factor that may reach sometimes 4, the latter having little impact in practice.I have also completed a genome-wide association study on photo-ageing. This study was performed on 502 Caucasian women characterized by their grade of photo-ageing, as measured by a well-established technology. In this study, the women were genotyped with OmniOne Illumina chips (1M SNPs), and I have identified two genes (STXBP5L et FBX040) associated with a SNP that passes the Bonferroni threshold, whose implication in photo-ageing was not suspected until now. Interestingly, this association has been highlighted with two other phenotypes which suggest a possible common molecular mechanism between sagging and wrinkling. There was no replication for the lentigin criteria, the third component studied of photo ageing.These studies are on the process to be published in international peer-reviewed scientific journals.
