Synteny is the evolutionary conservation of gene composition and collinearity between two genomes. We are conducting a global analysis of synteny between C. elegans and C. briggsae in an effort to understand the selective pressures that maintain or corrupt gene complement and genome structure. In particular, we are using synteny to define orthologs and paralogs, and are examining the extent and distribution of syntenic blocks and breaks in synteny within and between chromosomes.

We have utilized two complementary approaches to explore global synteny between C. elegans and C. briggsae. The first approach, at the gene level, uses sustained stretches of collinear ortholog pairs to delimit syntenic blocks. The second approach, at the nucleotide level, relies on sequence alignments generated with the WABA algorithm.

The ortholog-based approach to synteny begins with ab initio gene prediction, followed by assignment of ortholog pairs. Operationally, we conservatively assigned orthologs in two ways. First, we assigned genes to an ortholog pair if they were blastp best mutual matches and similar in length. A second approach (used by our collaborators Avril Coghlan and Marc Sohrmann) first "blessed" predicted C. briggsae exons that had a high degree of sequence identity, then assayed a combination of exon number, overall identity, and synteny between predicted genes and potential ortholog partners. The intersection of these two sets contains 7,378 primary ortholog pairs. Stretches of collinear orthologs were then used to define the limits of syntenic blocks and to score breaks in synteny. We are now expanding ortholog and paralog assignments using these syntenic blocks as a guide coupled with less-stringent metrics for identity than those used for the initial assignments. C. briggsae gene predictions and evidence for orthology will be available via WormBase (www.wormbase.org).

This approach to determining synteny rests heavily on both gene prediction and ortholog assignment algorithms, and will not reveal breaks in synteny contained entirely between two orthologs. To address these limitations, we also used a nucleotide-level similarity approach, exploiting sequence alignments generated using the WABA algorithm for cross species alignment of genomic sequence. Both approaches reveal a similar degree and extent of synteny between the two genomes, with between 60-75% of the C. elegans genome contained within syntenic blocks. Breaks in synteny are far more likely to be due to local, intrachromosomal events than interchromosomal events. Intrachromosomal breaks occur more frequently on autosomes than on the X chromosome, and more frequently on chromosome arms than within the gene-rich clusters. Interchromosomal breaks are more likely to occur between autosomes than between autosomes and the X chromosome. In exploring the context of these breaks, preliminary observations suggest that regions rich in repeats are less likely to be covered by syntenic blocks.

In addition to exploring the global synteny between C. elegans and C. briggsae, we are characterizing the selective pressure across the genome both on ortholog pairs and paralogs, as well as exploring the expansion and contraction of gene families.