Item

Abstract

Sequence comparison is an essential tool in modern biology. It is used to identify homologous regions between sequences, and to detect evolutionary relationships between organisms. Sequence comparison is usually based on alignments. However, aligning whole genomes is computationally difficult. As an alternative approach, alignment-free sequence comparison can be used. In my thesis, I concentrate on two problems that can be solved without alignment: (i) estimation of substitution rates between nucleotide sequences, and (ii) detection of local sequence homology. In the first part of my thesis, I developed and implemented a new algorithm for the efficient alignment-free computation of the number of nucleotide substitutions per site, and applied it to the analysis of large data sets of complete genomes. In the second part of my thesis, I developed and implemented a new algorithm for detecting matching regions between nucleotide sequences. I applied this solution to the classification of circulating recombinant forms of HIV, and to the analysis of bacterial genomes subject to horizontal gene transfer.