Protein structure analysis randomization

The greater number of folds in larger proteomes is intuitively obvious simply because the functioning of more complex organisms is expected to require a greater structural diversity of proteins. From a different perspective, the increase of diversity follows from a stochastic model, which describes a proteome as a finite sample from an infinite pool of proteins with a particular distribution of fold fractions ( a bag of proteins ). A previous random simulation analysis suggested that the stochastic model significantly (about twofold) underestimates the number of different folds in the proteomes (Wolf et al., 1999). In other words, the structural diversity of real proteomes does not seem to follow... 

The final session demonstrates how to characterize a protein region as random-izable. For a set of solvent-exposed residues of an immunoglobulin structure, 103 mutants are randomly generated. We examine how many of these mutants are destabilized with respect to the wild type. The analysis is repeated with an alternative set of residues that correspond to part of the natural epitope of the immunoglobulin structure. 

Advances in NMR instrumentation and methodology have now made it possible to determine site-specific proton chemical shift assignments for a large number of proteins and nucleic acids (1,2). It has been known for some time that in proteins the "structural" chemical shifts (the differences between the resonance positions in a protein and in a "random coil" polypeptide (3-5),) carry useful structural information. We have previously used a database of protein structures to compare shifts calculated from simple empirical models to those observed in solution (6). Here we demonstrate that a similar analysis appears promising for nucleic acids as well. Our conclusions are similar to those recently reported by Wijmenga et al (7),... 

Another historical method is the NOE cross-validation.47 This is performed by a partial elimination of NOEs from a full set of NOEs. The elimination is random. Calculated structures from the different sets of partially eliminated NOEs are statistically analyzed in comparison with the structure calculated from the full set of NOEs. The predictability of the eliminated NOEs from remaining NOEs is assessed as structural consistency (and therefore high quality) of the eliminated NOEs. With increased computing power, the statistical analysis will be feasible to determine high-resolution protein structure. 

We applied the neighbor-dependent sequence analysis on the residues of immediate neighbors in secondary structures of proteins (25,26). The NDPs are represented as (a i)j An sx(a 1 value of 1.0 means that the occurrence of the residue pair, ax (or xa), in the secondary structure j is the same as its frequency of occurrence in proteins. A value greater than 1.0 means the pair has occurrence in the secondary structure j higher than that in proteins, i.e., the pair has preference for adopting the j secondary structure conformation. A (a i), value lower than unity would suggest that the pair is less likely to adopt the j secondary structure than random distributions. For example, a eP A- ),a = 1-52 in short helices means that Pro has 62% more chance to be found in short helices than it in the proteins when it precede Ala, i.e., Pro at — 1 position of Ala. On the other hand, a ep(A+1)ia = 0.47 suggests that Pro is less likely to be found in short helix when it follows Ala, i.e., Pro at +1 position of Ala. 

Peptide CD structure analysis differs firom that of proteins as peptides are usually a combination of a particular secondary structure and random coil (i.e. unordered) residues. Proteins seldom if ever have random coils, so the CD of the proteins in the basis set of the protein anal3rsis program has no random coil component. Further, the analysis of the CD of a peptide is not under-determined, and so does not require a flexible method like Variable Selection. A Convex Constraint Analysis (22) that extracts the component spectra has been developed into a peptide structure analysis program by Greenfield (14). 

Shown is the fraction of an amino acid in a regular structural element, related to the fraction of all amino acids of the same structural element. P = 1 means random distribution P > 1 means enrichment, P < 1 means depletion. The data are based on an analysis of 66 protein structures. 

---