Home Programme

Wednesday, 8 December 2004
13:30 - 14:00

Phenyx: Achieving increased pertinence in protein identification and characterization

Pierre-Alain Binz^1,2, Alexandre Masselot^1
1GeneBio SA, Geneva, Switzerland; ²Swiss Institute of Bioinformatics and Geneva University, Geneva, Switzerland

Mass spectrometry is a central analytical tool in many experimental workflows in proteomics. Protein identification based on MS data is an essential component of the data interpretation pipeline. Many bioinformatics tools are available that aim to most efficiently extract the best of the MS data to help end-users in deciphering the protein content of analysed samples. Nevertheless, many end-users are still faced with the problem of handling and comfortably validating huge amounts of data while trying to associate the obtained results with data quality and biological pertinence.

Here we will present Phenyx, the result of close collaboration between wet-lab and dry-lab proteomics scientists. Phenyx is not just another protein identification tool; it is a complete working environment.

Phenyx has a client-server architecture. It consists of a core calculation engine (PCE) on one end that performs the identification and manages jobs, user profiles, user data and results. On the other end, the communication with the calculation unit is handled either directly via scripts, allowing easy and seamless integration in automated pipelines, or via the Phenyx Virtual Desktop, a java application.

The core Phenyx Calculation Engine (PCE) incorporates the true statistical scoring models OLAV developed at GeneProt Inc. for MS/MS data interpretation. Each scoring scheme can be fine-tuned for all individual experimental and instrumental MS/MS setups. Therefore, as examples, specific scoring schemes can be calculated for different values of the voltages in an ion-trap analyser; distinct schemes can be defined for samples issued from an isotope-labeling or a MuDPIT experiment or for a simpler workflow implying digested spots from 2-DE gels. For the specific needs of PMF calculation, Phenyx incorporates Aldente in collaboration with the Swiss Institute of Bioinformatics. Both approaches are able to query the detailed annotations of the UniProt-SwissProt database, and therefore search for described alternative splicing events, mutations or post-translational modifications. It is available as a public Linux server on Vital-IT or as a monoprocessor system that can be installed on a PC (see www.phenyx-ms.com/downloads/downloads.html).

The main client is the Phenyx Virtual Desktop (PVD), a java application designed to manage job submission, progress visualization and dynamic results evaluation. End-users have access to their own profiles, sets of parameters (including “private” enzymatic rules and amino acid modifications) and can manage their own identification jobs. Comparison of results can be easily performed and all tables can be exported into Excel. Results of quantitation experiments involving technologies such as iTRAQ can be analysed.

Evaluation of reproducibility and measure of differences between technologies are available through the job comparison feature. Identically identified proteins (and possibly peptides) can be simultaneously validated and exported into a single report.

In terms of quality and pertinence assessment of the processed data and interpreted results, the stability of an MS instrument and the global pertinence of the identified peptides can be observed via statistics graphs. Detailed interpretations of peptide and protein assignments are accessible via dynamic graphics and tables.

Congress management: