What do different chromosomes have in common? An exhaustive but efficient computation

VERÓNICA BECHER

Barcelona, España

Otro; Seminar Center for Genomic Regulation; 2009

Center for Genomic Regulation,

What do different chromosomes have in common? An exhaustive but efficient computationEukaryote genomes are about 95% non-coding sequences. Traditionally considered unuseful, this dark matter of the genomes contain elements of variable sizes with notable evidences of regulatory functions. Much effort has been done to characterize the repetitive sequences found in the human genome and those belonging to other species. Nevertheless, a quantitative analysis of the dynamics of repetitive elements is still missing.We aim at studying the number and the distribution of repetitive DNA sequences  in the set of chromosomes of 10 species of mammals. The questions we pose are related to the frequency and distribution of those elements explaining the similarities between chromosomes in all-against-all comparisons, within and between species.As a preliminary phase we made an exhaustive pairwise chromosome comparison of the Homo Sapiens genome. We carried out the highly demanding computational task of exhaustively finding all repetitive sequences larger than 40bp occurring simultaneously in any two chromosomes, with exact matching. Computationally this is a hard problem with respect to time and memory since, for instance, the HS chromosome 1 has 250 million nucleotides, and its repetitive sequences make up roughly 50% of it. Known algorithms did not scale up to efficiently deal with two entire chromosomes simultaneously.We quantified the  sequences found, gave statistics, made biological identifications, and provided various insightful graphical visualizations. We also give a notion of quantity of commom information between chromosomes, and a distance function that measures how much additional information is needed to obtain one chromosome given the other. We computed a matrix with the distance between every two chromosomes of 5 mammalian species: Homo sapiens, Pan troglodytes, Mus musculus, Rattus norvergicus and Cannis familiaris. Our current results reveal more information in common among chromosomes within species, rather than between species. In particular, each Homo Sapiens chromosome shares approximately 10%  with every other human chromosome. This 10% is distributed more or less evenly  along the chromosome surfaces, and it is mostly made of L1/LINE transposons. In this talk I will summarize the results obtained in this ongoing project.This experimental first phase is available online from http://kapow.dc.uba.ar/pats.