Similarity between related genomes may carry information on selective constraint in each of them.
We analysed patterns of similarity between several homologous regions of Caenorhabditis elegans
and C. briggsae genomes. All homologous exons are quite similar. Alignments of introns and of
intergenic sequences contain long gaps, segments where similarity is low and close to that between
random sequences aligned using the same parameters, and segments of high similarity.
Conservative estimates of the fractions of selectively constrained nucleotides are 72%, 17% and
18% for exons, introns and intergenic sequences, respectively. This implies that the total number
of constrained nucleotides within non-coding sequences is comparable to that within coding
sequences, so that at least one-third of nucleotides in C. elegans and C. briggsae genomes are
under strong stabilizing selection.