This SBIR project proposes the development of a software package containing bio-computational tools to facilitate accurate diagnosis of cancer based on classification of proteomics data obtained through mass spectrometry (MS). The software will accept MS data produced from a wide range of instruments but is specifically targeted toward MALDI/SELDI TOF analysis (Applied Biosystems, Ciphergen). The Phase I work will focus on exploring the applicability of various existing and modified statistical approaches for signal conditioning (Wiener filters) and classification (linear discriminant analysis, principal components analysis, support vector machines, Bayes method) of SELDI data derived from analysis of sera from individuals with pediatric Hodgkin's disease and acute myeloid leukemia. A software package to perform this task does not currently exist; therefore, the proposed research has significant potential for technical innovation. Cross-validation will be used to obtain an unbiased estimate of the performances of the classifiers. The TRIFT II equipment at the ARC (W&M) will be used to provide high-resolution MS TOF SIMS data to calibrate the SELDI equipment at EVMS. The Phase II project will leverage the research and proof-of-concept tools developed in Phase I to produce a commercial software package that will be licensed to researchers, as well as equipment manufacturers for inclusion with their instruments.
| Malyarenko, Dariya I; Cooke, William E; Tracy, Eugene R et al. (2006) Deconvolution filters to enhance resolution of dense time-of-flight survey spectra in the time-lag optimization range. Rapid Commun Mass Spectrom 20:1661-9 |
