Bioinformatics analysis identify novel OB fold protein coding genes in C. elegans

Abstract

Background: The C. elegans genome has been extensively annotated by the WormBase consortium that uses state of the art bioinformatics pipelines, functional genomics and manual curation approaches. As a result, the identification of novel genes in silico in this model organism is becoming more challenging requiring new approaches. The Oligonucleotide-oligosaccharide binding (OB) fold is a highly divergent protein family, in which protein sequences, in spite of having the same fold, share very little sequence identity (5-25%). Therefore, evidence from sequence-based annotation may not be sufficient to identify all the members of this family. In C. elegans, the number of OB-fold proteins reported is remarkably low (n=46) compared to other evolutionary-related eukaryotes, such as yeast S. cerevisiae (n=344) or fruit fly D. melanogaster (n=84). Gene loss during evolution or differences in the level of annotation for this protein family, may explain these discrepancies.

Methodology/principal findings: This study examines the possibility that novel OB-fold coding genes exist in the worm. We developed a bioinformatics approach that uses the most sensitive sequence-sequence, sequence-profile and profile-profile similarity search methods followed by 3D-structure prediction as a filtering step to eliminate false positive candidate sequences. We have predicted 18 coding genes containing the OB-fold that have remarkably partially been characterized in C. elegans.

Conclusions/significance: This study raises the possibility that the annotation of highly divergent protein fold families can be improved in C. elegans. Similar strategies could be implemented for large scale analysis by the WormBase consortium when novel versions of the genome sequence of C. elegans, or other evolutionary related species are being released. This approach is of general interest to the scientific community since it can be used to annotate any genome.

Figure 1. Superimposition of the novel OB-fold 3D-model with their templates.

(Light blue): Predicted 3D-models, (Wheat) PDB template. (.XXXX.)-nxxx name correspond to the protein name followed by the PDB code of the template.

Figure 2. Discovery Pipeline of novel OB fold protein coding genes.

It contains 3 Discovery Modules. SeqDIM: Sequence alignment DIscovery Module; StrucDIM:3D Structure prediction Discovery Module; and a Functional prediction Discovery Module FuncDIM.