Large scale proteomics data analysis

Label-free nanoLC-MS/MS workflows aim to translate sets of tandem mass spectrometry (MS/MS) spectra (usually of fractionated proteomes) to peptide and protein identification and quantification and to link proteome data to patient samples, clinical and other molecular data. We aim to identify proteins with regulated levels to enable discovery of candidate biomarkers or drug targets. Biomarker candidates from discovery experiments are followed-up by SRM in larger clinical sample collections (eg., serum, stool, CSF). Candidate biomarker validation by proteomics requires targeted mass spectrometry. A large number of mass spectrometry datasets have been generated at the OncoProteomics laboratory at the VU medical center. One of the main challenges is to select from a vast pool of candidate markers a small subset to follow up in the validation phase. Furthermore, candidate markers might not be specific to the disease under study. It has been observed that biomarkers found several studies of different diseases overlap. Hence our ambition is to analyze every dataset in the context of a large number of related experiments. Ideally, the data should be processed (and reprocessed) in highly similar if not identical workflows. In addition, the OncoProteomics laboratory is participating in the CTMM TraIT project where in collaboration with NBIC we aim to offers easy connections between -omics platforms and other clinical data to enable complex integrative data analysis. There is a need to store raw and processed data in a reliable and easily accessible manner. Currently, we have exceeded our local storage capacity for raw data. Next generation mass spectrometers that we are setting up will scale the data production by an order of magnitude. We request for support for reliable data storage of 2 Terabytes of existing data now and approximately 5 to 10 Terabytes per year in the coming 4 years. In addition, we wish to request for processing power to experiment with grid computing infrastructure to support our ambition of data mining at large scale, including the possibility of reanalyzing of legacy mass spectrometry data with newly developed computational tools.