Elements of coevolution in biological sequences

Studies of coevolution of amino acids within and between proteins have revealed two types of coevolving units: coevolving contacts, which are pairs of amino acids distant along the sequence but in contact in the three-dimensional structure, and sectors, which are larger groups of structurally connected amino acids that underlie the biochemical properties of proteins. By reconciling two approaches for analyzing correlations in multiple sequence alignments, we link these two findings together and with coevolving units of intermediate size, called "sectons," which are shown to provide additional information. By extending the analysis to the co-occurrence of orthologous genes in bacterial genomes, we also show that the methods and results are general and relevant beyond protein structures.