Dropout alignment allows homology recognition and evolutionary analysis of rDNA intergenic spacers

Subrepeats within the ribosomal gene (rDNA) intergenic spacer (IGS) play an important role in enhancing RNA polymerase I transcription. Despite this functional role and presumed selective constraint, there is surprisingly little sequence similarity among IGS subrepeats of different species. This sequence dissimilarity corresponds with the fast insertion-deletion (indel) rates observed in short mononucleotide microsatellites (here referred to as poly[N] runs, where N is any nucleotide), which are relatively abundant in rDNA IGS subrepeats. Some species have different types of IGS subrepeats that share species-specific poly(N) run patterns. This finding indicates that many IGS subrepeats within species share a common evolutionary history. Furthermore, by aligning sequences after modifying them by the dropout method, i.e., by disregarding poly(N) runs during the sequence aligning step, we sought to uncover evolutionarily shared similarities that fail to be recognized by current alignment programs. To ensure that the improved similarities in the computed alignments are not a chance artifact, we calibrated and corrected the IGS subrepeat sequences for the influence of repeat length and estimated the statistical significance of the alignments (in terms of a stringent p-value) obtained by the dropout method by comparing them to null models constructed using random sequence sets from the same genomes. We found that most diverse kinds of rDNA IGS subrepeats in one species must have been derived from a common ancestral subrepeat, and that it is possible to infer the evolutionary relationships among the IGS subrepeats of different species by comparative genomics methods based on dropout alignments.