Motivation: Methods that predict the structure of molecules by looking for
statistical correlation have been quite effective. Unfortunately, these met
hods often disregard phylogenetic information in the sequences they analyze
. Here, we present a number of statistics for RNA molecular-structure predi
ction. Besides common pair-wise comparisons, we consider a few reasonable s
tatistics for base-triple predictions, and present an elaborate analysis of
these methods. All these statistics incorporate phylogenetic relationships
of the sequences in the analysis to varying degrees, and the different nat
ure of these tests gives a wide choice of statistical tools for RNA structu
re prediction.
Results: Starting from statistics that incorporate phylogenetic information
only as independent sequence evolution models for each position of a multi
ple alignment, and extending this idea to a joint evolution model of two po
sitions, we enhance the usual purely statistical methods (e.g. methods base
d on the Mutual Information statistic) with the use of phylogenetic informa
tion available in the sequences. In particular; we present a joint model ba
sed on the HKY evolution model, and consequently a chi(2) test of independe
nce for two positions. A significant part of this work is devoted to some m
athematical analysis of these methods. We tested these statistics on region
s of 16S and 23S rRNA, and tRNA.