Two approaches for improving the phylogenetic analysis of ribosomal RNA sequences are presented, both within a parsimony framework. The first deals with characterizing difficult-to-align regions based on secondary structural properties. This approach uses structure in conserved areas of a multiple sequence alignment to isolate 1. regions of ambiguous alignment (RAA), 2. regions of slipped-strand compensation (RSC), and regions of helix expansion and contraction (REC). This dissection of ambiguously-aligned regions into discrete classes provides an objective means for assigning nucleotide positional homology, and also retains phylogenetic information via fragment level alignment methods (i.e., INAASE, FSO).

The second method deals with characterizing base-pairs within conserved helices as single states, providing a 20-state approach to analyzing RNA datsets (4 states for unpaired bases, 16 states for paired bases). This method allows for differential weights to be assigned to all states, either paired or unpaired, providing an alternative to previously proposed "stem-loop" weighting schemes. Perl scripts designed to convert structurally-aligned RNA datasets into this 20-state format are discussed.