Computational biology has focused primarily on globular proteins; this application proposes to begin addressing questions related to fibrous proteins using computational analysis on human disease and polymorphism databases. Recent studies suggested it may be possible to predict on the basis of molecular structure and sequence conservation when an amino acid replacement will be deleterious and lead to a clinical phenotype. Computational analyses is proposed on missense mutations in collagen and coiled coil domains, two of the major motifs found in fibrous proteins. It is hypothesized that the repeating sequence patterns and superhelical structures of these motifs will result in biasing of the observed mutations in terms of location within the motif and the identity of the substituting residue, and that the location of mutations within the entire repeating domain may relate to higher order structures and binding sites. The set of observed mutations will be compared with the entire assembly of amino acid replacements expected on the basis of the nucleotide substitution rates. The possibility that variations in collagen sequence could confer a propensity to common diseases such as osteoporosis and the set of all observed non-synonymous single nucleotide polymorphisms will also be examined.
