Inferring RNA Structure by Phylogenetic and Genetic Analyses

Abstract

Comparative sequence analysis (CSA) is the process in which related macromolecular sequences are collected and compared in order to extract information about common underlying structures. The idea of inferring structure by comparing sequences is rooted in the well-verified fact that the architecture of macromolecules evolves much more slowly than their sequence. As a consequence, nature usually presents us with a significant fraction of all possible sequences from which particular structures or substructures may be generated and, in so doing, tempts us to guess what these structures might be. The sequences to be fed to CSA thus used to be the products of natural selection, but, as discussed in Baskerville et al. (this volume), they should increasingly come in the future from experiments in selection and evolution devised by humans. However, even though the source of variation is changing, the way CSA is carried out, and its limitations, should remain largely the same.

The classic example of a structure successfully predicted by CSA is the tRNA cloverleaf (Madison et al. 1966; RajBhandary et al. 1966; Zachau et al. 1966), but in fact, the Watson-Crick pairs themselves were inferred by applying the very principles of comparative analysis—invariance of higher-order structure—to possible pairings between the hydrogen-bond donor and acceptor groups of the bases (Watson and Crick 1953). CSA could, in principle, make use of any type of macromolecular sequences, and there are examples of CSA being successfully utilized to deduce interactions between proteins and nucleic acids (see, e.g., Nardelli et al....