The interplay of proteomics, genomics and bioinformatics approaches and their potential for cancer diagnosis and prognosis

Parkinson, Erika, Matharoo-Ball, Balwir, Ball, Graham, Shardendorf, Dirk, Ugurel, Selma, Creaser, Colin and Rees, Robert C. (2006) The interplay of proteomics, genomics and bioinformatics approaches and their potential for cancer diagnosis and prognosis. AACR Meeting Abstracts, 47, 687.


Full text not available from this repository.


To gain a comprehensive understanding of the physiology and pathophysiology of cancer an approach that harmoniously integrates the various omic' platforms is key to cancer biomarker discovery. We have used a combination of high throughput protein pattern detection methods using matrix assisted mass spectrometry time-of-flight (MALDI-TOF) instrumentation and in some studies with protein chip technology to investigate discriminatory protein patterns in melanoma patients in matched serum and plasma and primary tumor cell lines. The cell lines have further been studied using a genomic based method of RT-PCR to give identity to the expressed genes at the time of tumor excision. The gene mutations studied were BRAF, P16, TP53, PTEN, NRAS, INK4A, CTNNB1, and CDK4. Artificial neural networks (ANNs) and descriptive statistics were applied to the combined proteomic and genomic data for the cell lines and protein patterns for matched serum and plasma to identify discriminatory patterns with different clinical disease states in melanoma and to further identify the important biomarkers for the future diagnosis and prognosis of this cancer. Preliminary results for the protein fingerprint patterns and a TP53 gene mutation in metastatic melanoma cell lines showed that the ANNs were capable of predicting with 99% confidence in a blind sample set whether the cell line had a gene mutation or not. For the serum melanoma study the ANNS using proteomic "fingerprint" identified, 9 ions to date. The 9 ion ANNs model classified the data correctly with a median accuracy of 92.3 % (inter-quartile range 89.4 - 94.9 %) for a separate test set of data set aside for validation over 50 random sample cross validation data splits. All ions show statistically significant increase/decrease in intensities. Some peaks could be identified by eye, some cannot

Item Type: Article
ISSNs: 1948-3279
Related URLs:
Subjects: R Medicine > RC Internal medicine > RC0254 Neoplasms. Tumors. Oncology (including Cancer)
Divisions : Faculty of Natural and Environmental Sciences > Biological Sciences > Molecular & Cellular
ePrint ID: 340252
Accepted Date and Publication Date:
Date Deposited: 15 Aug 2012 14:13
Last Modified: 31 Mar 2016 14:29

Actions (login required)

View Item View Item