Data base for protein sequences

Abstract

Proteins are linear polymers synthesized in living organisms from twenty different kinds of amino acids according to the message carried in the chromosomes. Typically, they have evolved by natural selection over hundreds of millions of years through many small changes in sequence. The present computerized data base includes 77,267 amino acid residues from 767 sequences. We believe that all of the protein structures occurring in living organisms can be combined into fewer than 1,000 groups containing proteins of similar sequence. Each group can be characterized by a few sequences that are known exactly and by a number of evolutionary parameters. The rest of the structures, occurring in organisms not examined, can be described with estimated precision in terms of the number of differences from a known sequence or from sequences inferred to have been present in ancestral forms. The following kinds of information are needed: sequences of proteins from each group, the phylogenetic tree of biological species, a list of protein groups and the gene duplications in each, a description of the quantitative parameters of the evolutionary processes affecting proteins, and methods of estimation of sequences in ancestral species and in living forms phylogenetically close to those investigated. The conceptual tools and the computer programs necessary for the prediction of all of the 1010 to 1011 protein sequences in living species are described. One can readily visualize the separate parts operating as an integrated interactive computerized data base that could predict sequences for specified organisms with an estimated precision based on the collection of known sequences.