A new software pipeline for analysis of high-performance mass spectrometer data allows rapid and accurate identification of potential biomarkers for detection of diseases such as cancer or other patient conditions. The software exploits the mass precision and resolution of high-performance instrumentation, bypassing peak finding steps and instead using discrete m/z data points to identify putative biomarkers. The technique is insensitive to peak shape and works well on overlapping and non-Gaussian peaks which can confound peak-finding algorithms. Software modules use data from known samples to identify differences between healthy and diseased patients, a training phase, then these differences may be used to classify patients whose conditions are unknown, a testing phase. Alternatively, the differences identified in the training phase may be analyzed to find the underlying biochemical species, which can then be tested as a biochemical marker for the disease or condition. The method has been demonstrated on samples from patients with ovarian cancer, prostate cancer, multiple sclerosis, Alzheimer’s disease, and Parkinson’s disease.

Details of the method are published in the Journal of Bioinformatics and Computational Biology, vol. 5, no. 5 (2007) pp. 1023-1045.

(click pictures for larger format)