Completion of autobuilt protein models using a database of protein fragments

Abstract

Two developments in the process of automated protein model building in the Buccaneer software are presented. A general-purpose library for protein fragments of arbitrary size is described, with a highly optimized search method allowing the use of a larger database than in previous work. The problem of assembling an autobuilt model into complete chains is discussed. This involves the assembly of disconnected chain fragments into complete molecules and the use of the database of protein fragments in improving the model completeness. Assembly of fragments into molecules is a standard step in existing model-building software, but the methods have not received detailed discussion in the literature.

Figure 1

Running distance-matrix representation of a single fragment, where D _ij is the distance between the ith and jth C^α atoms. The shaded cells are those available for loop fitting using only two C^α atoms at each end of the fragment.

Figure 2

Running distance-matrix representation of the protein-chain database, where D _i,j is the distance between the ith and jth C^α atoms. The shaded cells are those which would be used to score the fit of a search fragment against a particular range in the database.

Figure 3

Tail plot of the proportion of search fragments for which the fit of the best-matching fragment is worse than a given criterion for different fragment lengths. (a) R.m.s. deviation of C^α positions between the best database fragment and the search fragment; (b) maximum deviation of any C^α positions between the best database fragment and the search fragment.

Figure 4

Partially incorrect assembly of the model for 1vlu from multiple fragments. The wrongly positioned region is shown in black (a) in the Buccaneer model and (b) in the deposited structure.