Evaluating cancer protein identification from mass spectroscopy data

Early detection of ovarian cancer substantially increases patient survivability. Mass spectrometry data sets for the purpose of research into early detection methods have been funded by the National Ovarian Cancer Early Detection Program, gynecologic oncology clinic at Northwestern University Chicago IL, USA. Researchers have applied different machine learning methods for classification on several of the spectra data sets, (including the most recent produced on an ABI Hybrid Pulsar QqTOF instrument (Q-Star) mass spectrometer) for the purpose of predicting cancer. While identification of the proteins or metabolites within these spectra is not required for early detection, it would give a better understanding of the disease processes. More importantly, the potential for discovery of new drug targets would be possible by identification of cancer diagnostic proteins.;The aim of this research is to employ bioinformatics techniques to identify proteins associated with peaks of interest, based purely on their mass values. These proteins of interest would then be targeted for further research. This is a new paradigm for the discovery of cancer proteins that researchers can use in their investigation of potential biomarkers and drug targets. This approach brings together data-mining and statistical analysis in a high-throughput data-analysis-pipeline approach for knowledge discovery. Statistical validation of data mining results of keyword searches has shown that differences in cancer relationships exist between proteins discovered by classification and those collected randomly. The final product of this work is the delivery to the cancer research community known and previously unknown cancer proteins with corresponding probabilities of disease significance.