DNA fragments positioning improvement to realise consensus sequences in the de novo long reads assembly context
Year: 2020
Type: dissertation
Abstract: \n DNA molecular in silico analysis requires sequencing it in fragments called reads, and then assembling them. Today’s, long reads sequencing technologies offer the possibility to overcome genome’s repeated regions issues with being entirely covered, but product high errors rate data with sequencing errors, like nucleotides insertions or deletions,called indels. De novo assembly is an assembly without using a reference. Although some assemblers already exist according to several methods - as using De Bruijn graphs or by correcting iteratively the reads for example, we propose a two steps strategy : first, we attribute to the maximum of reads a position on a same positions axis and then weproduct a consensus sequence thanks to the positioning. At these aims, we propose a modelling for the positioning issue with the mixed integer linear programming (MILP), and we present first ideas for the consensus sequences production with MILP too and multiple sequences alignment from the positioning. As the final aim of this strategy isto formalize the genome assembly issue, we structured it according the mathematical method, that permits to target methodological choices precisely, and then reducing the heuristic uses. Finally, we tested the strategy with bacteria genomes. Despite the fact that positioning results show positive ones, the consensus results are less positive but don’tremove the potentiality of the associated methods.\n
Source: HAL
Author: Epain, Victor
Language: English
Subfield: Molecular Biology
Domain: Life Sciences
Open Access status: green
OpenAlex ID: W4287756331