Prediction of partial membrane protein topologies using a consensus approach

Abstract

We have developed a method to reliably identify partial membrane protein topologies using the consensus of five topology prediction methods. When evaluated on a test set of experimentally characterized proteins, we find that approximately 90% of the partial consensus topologies are correctly predicted in membrane proteins from prokaryotic as well as eukaryotic organisms. Whole-genome analysis reveals that a reliable partial consensus topology can be predicted for approximately 70% of all membrane proteins in a typical bacterial genome and for approximately 55% of all membrane proteins in a typical eukaryotic genome. The average fraction of sequence length covered by a partial consensus topology is 44% for the prokaryotic proteins and 17% for the eukaryotic proteins in our test set, and similar numbers are found when the algorithm is applied to whole genomes. Reliably predicted partial topologies may simplify experimental determinations of membrane protein topology.

Fig. 1.

Fraction correctly predicted global topologies (black bars) and fraction of the test sets covered (white bars) for different levels of agreement among the five prediction methods (5/0, all methods agree; 4/1, four methods agree, etc.). Left: prokaryotic proteins; right: eukaryotic proteins.

Fig. 2.

Partial consensus topology prediction procedure (for simplicity shown only for two prediction methods). (A) For each residue position, a consensus topology state (i, o, m, w) is assigned only if all five prediction methods agree. Otherwise, no consensus is assigned (designated as ‘.’). If two loops with opposite locations (o and i) appear at the same position, a loop clash (X) is assigned. In the same manner, a TMH clash (#) is assigned when two TMHs with opposite orientations (m and w) appear at the same position. After the filtering step described in B, the consensus topology is scanned from the N–terminus, and the beginning of the first PCT is defined by the first TM region (m or w states) of at least n consecutive residues in the consensus topology (where the default value is n = 5). The PCT is then extended towards the C–terminus until either a consensus TMH of less than n consecutive residues is encountered, or a loop– or TM–clash occurs. In either case, the end of the PCT is defined by the most C–terminally located TMH in the consensus topology. The process of PCT construction is then repeated until the C-terminal end of the protein is reached. (B) Removal of TMH clashes which are due to slight misalignments of the TMHs. When a TMH clash occurs between two partially overlapping consensus TMHs with opposite orientations, the clash state is replaced by a loop state of the appropriate type. To emphasize that the resulting loop was obtained in the filtering procedure (and not in the majority–vote procedure), it is indicated in lowercase. (C) The figure illustrates how the choice of n–value affects the PCT prediction results. The result can differ substantially when a consensus TMH that is longer (left) or shorter (right) than n is encountered. (D) Average fraction of sequence length covered by PCTs (squares) and average fraction of correctly predicted PCTs (circles) for different values of n for the prokaryotic (black symbols) and eukaryotic (white symbols) test sets.