Building Genomic Profiles for Uncovering Segmental Homology in the Twilight Zone

Abstract

The identification of homologous regions within and between genomes is an essential prerequisite for studying genome structure and evolution. Different methods already exist that allow detecting homologous regions in an automated manner. These methods are based either on finding sequence similarities at the DNA level or on identifying chromosomal regions showing conservation of gene order and content. Especially the latter approach has proven useful for detecting homology between highly divergent chromosomal regions. However, until now, such map-based approaches required that candidate homologous regions show significant collinearity with other segments to be considered as being homologous. Here, we present a novel method that creates profiles combining the gene order and content information of multiple mutually homologous genomic segments. These profiles can be used to scan one or more genomes to detect segments that show significant collinearity with the entire profile but not necessarily with individual segments. When applying this new method to the combined genomes of Arabidopsis and rice, we find additional evidence for ancient duplication events in the rice genome.