Metagenomic data analysis : computational methods and applications

Metagenomics is the study of the genomic content of microbial communities, acquired through DNA sequencing technology. The main advantage of metagenomics is that it can overcome the limitations of individual genome sequencing, that can work only on the few culturable microbes. Unfortunately, the analysis of metagenomic data is challenging. Firstly, we do not know from which genome a DNA sequence of the data, called read, was sampled. In most of the cases, the full sequences of the source genomes are not available, and even their number is unknown. Secondly, the information contained in a metagenome is scattered among reads much shorter then their source genomes. The topic of this thesis is the computational analysis of metagenomic data in order to understand their content. We focus mainly on binning, that consists in clustering the reads according to the biological information they contain. Chapters 2 and 3 present two binning methods. The former clusters reads according to their similarity with proteins, using combinatorial optimization. The second performs taxonomic assignment of reads by clustering reads at multiple taxonomic ranks. Chapter 4 deals with genomic signatures, which are functions frequently adopted by binning methods to have a vector space representation of the reads. We compare experimentally the performances of different genomic signatures on metagenomic data; new signatures are also studied. A case study in metagenomics data annotation is described in Chapters 5 and 6. The related metagenomic project investigates the anammox bacteria Candidatus `Brocadia fulgida', sequencing the same microbial community with three different technologies. Anammox bacteria have been successfully employed to treat highly loaded wastewater in industrial and municipal wastewater treatment systems. In Chapter 5, in particular, we try to retrieve the core genes of anammox metabolism. Chapter 6 focuses on a comparative quantitative analysis of the metagenomes acquired by the three technologies