U OMASITS¹, AR VARADARAJAN¹, DJ STEKHOVEN², M QUEBATTE³, M ROBINSON², C DEHIO³, CH AHRENS<sup>2

1</sup>Agroscope, Institute for Plant Production Sciences, 8820 Wädenswil, Switzerland, ²SIB, Swiss Institute of Bioinformatics, 8057 Zurich, Switzerland, ³Biozentrum, Infection Biology Unit, University of Basel, 4056 Basel, Switzerland

Large advances in functional genomics technologies have enabled the description of complete transcriptomes with RNA-Seq. In contrast, the study of complete proteomes has largely lagged behind. We have previously devised a generic strategy to describe complete, condition-specific expressed prokaryotic proteomes using Bartonella henselae as a model system [1]. Applying this strategy under two conditions that mimic interaction of B. henselae with its major hosts led to the description of a - to our knowledge for the first time - complete membrane proteome, including expression evidence for all members of the VirB/D4 type IV secretion system (T4SS).


This unique dataset was further exploited to determine the predominant subcellular localization of more than 90% of the expressed proteins [2]. Importantly, the high quality proteomics data from this large dataset provided evidence for the expression of a number of short, unannotated ORFs. As the annotations of different reference genome annotations can largely differ, we have developed a generic proteogenomics strategy, that allows researchers to discover the full coding potential of prokaryotic organisms. So far, we were able to uncover missed protein-coding ORFs in each prokaryote tested.


We discuss the manifold implications of this generic approach, up to the identification of coding differences from clinical genomes sequenced and assembled from long PacBio reads, a capability urgently needed for personalized medicine.


[1]. Omasits U et al., Genome Research 2013, 23:1916-1927.
[2]. Stekhoven DJ, Omasits U et al., J Proteomics 2014, 99:123-137.