Brad Gulko and David Haussler . Using Multiple Alignments and Phylogenetic Trees to Detect RNA Secondary Structure [490k postscript file]
We describe a statistical method to determine if a pair of columns in a multiple alignment of a homologous family of RNA sequences shows evidence of being base paired. The method makes explicit use of a given phylogenetic tree for the sequences in the alignment. It is tested on a multiple alignment of 16S rRNA sequences with good results.
Jesse Reklaw, Rebecca Hanna and David Haussler . Prediction of ribosomal RNA base-pairing by neural networks
With the large number of rRNA sequences available in multiple alignments, comparative sequence analysis has enabled the reliable inference of higher order structure. This paper presents a neural net method for predicting base-pairing in the SSU and LSU of rRNA. Input is coded as the observed counts between pairs of bases and an interdependence measure known as mutual information. Experiments conducted on several data sets evaluate the network's overall performance and its ability to generalize across domains and macromolecules. With a large multiple alignment representative of the phylogenetic diversity, our best method characterizes over 96% of the test data as base-pairing or non-base-pairing. This accuracy may be sufficient to predict some tertiary interactions, although the rate of false positives still limits the applicability of the method. The issue of generalization to other macromolecules also remains unresolved.