GibbsCluster: unsupervised clustering and alignment of peptide sequences

Abstract

Receptor interactions with short linear peptide fragments (ligands) are at the base of many biological signaling processes. Conserved and information-rich amino acid patterns, commonly called sequence motifs, shape and regulate these interactions. Because of the properties of a receptor-ligand system or of the assay used to interrogate it, experimental data often contain multiple sequence motifs. GibbsCluster is a powerful tool for unsupervised motif discovery because it can simultaneously cluster and align peptide data. The GibbsCluster 2.0 presented here is an improved version incorporating insertion and deletions accounting for variations in motif length in the peptide input. In basic terms, the program takes as input a set of peptide sequences and clusters them into meaningful groups. It returns the optimal number of clusters it identified, together with the sequence alignment and sequence motif characterizing each cluster. Several parameters are available to customize cluster analysis, including adjustable penalties for small clusters and overlapping groups and a trash cluster to remove outliers. As an example application, we used the server to deconvolute multiple specificities in large-scale peptidome data generated by mass spectrometry. The server is available at http://www.cbs.dtu.dk/services/GibbsCluster-2.0.

Figure 1.

Clustering results for the Fibroblast dataset. The solution with highest KLD consists of four clusters, and the corresponding sequence motifs are shown as sequence logos.

Figure 2.

Comparison of unsupervised clustering to HLA restrictions assigned by NetMHCpan on Fibroblast data. Left: distribution of percentile rank scores predicted by NetMHCpan for the allele dominating each cluster; for the trash cluster, the best predicted rank score to any of the six alleles was used. Right: sequence logos from literature (made with MHCcluster (16)) of the alleles found in each cluster; group 2 is composed mostly of ligands predicted by NetMHCpan to be restricted to three different alleles with similar binding motifs.

Figure 3.

Length profile of peptides in the optimal Fibroblast clustering solution. Solid lines represent, for each group, the percentage of peptides with a given length over the total number of peptides in the group. The stacked bar plot in the background is the corresponding length frequency (number of ligands of a given length in a given group divided by the total number of peptides of that length in all groups).