Assessing the state of substitution models describing noncoding RNA evolution.

摘要: Phylogenetic inference is widely used to investigate the relationships between homologous sequences. RNA molecules have played a key role in these studies because they are present throughout life and tend evolve slowly. has been shown be dependent on substitution model used. A wide range of models developed describe evolution, either with 16 states describing all possible canonical base pairs or 7 where 10 mismatched nucleotides reduced single state. Formal selection become standard practice for choosing an inferential works well comparing specific type, such as comparisons within nucleotide amino acid models. Model cannot function across different sized state spaces likelihoods conditioned data. Here, we introduce statistical state-space projection methods that allow direct comparison 7-state 16-state To demonstrate general applicability our new methods, extract 287 families from genomic alignments perform selection. We find 281/287 families, selected preference models, simple more conserved shorter stems complex divergent longer stems. Other factors, molecule GC-content, limited impact Our freely available open-source PHASE 3.0 software.