Developing a Discrimination Rule between Breast Cancer Patients and Controls Using Proteomics Mass Spectrometric Data: A Three-Step Approach

Abstract

To discriminate between breast cancer patients and controls, we used a three-step approach to obtain our decision rule. First, we ranked the mass/charge values using random forests, because it generates importance indices that take possible interactions into account. We observed that the top ranked variables consisted of highly correlated contiguous mass/charge values, which were grouped in the second step into new variables. Finally, these newly created variables were used as predictors to find a suitable discrimination rule. In this last step, we compared three different methods, namely Classification and Regression Tree (CART), logistic regression and penalized logistic regression. Logistic regression and penalized logistic regression performed equally well and both had a higher classification accuracy than CART. The model obtained with penalized logistic regression was chosen as we hypothesized that this model would provide a better classification accuracy in the validation set. The solution had a good performance on the training set with a classification accuracy of 86.3%, and a sensitivity and specificity of 86.8% and 85.7%, respectively.

Mass spectrometry of peptides and proteins from human blood

“…Statistical analysis might be applied to proteomic data at several levels from power calculations and experimental design , peptide patterns (Ragazzi et al, 2006;Ressom et al, 2006;Au et al, 2007;Heidema & Nagelkerke, 2008;Helleman et al, 2008), or parent peptide intensity analysis (Marshall et al, 2003). Statistical test such as t-tests (Negishi et al, 2009), ANOVA (Marshall et al, 2003;Higgs et al, 2008;Zhao et al, 2008a;Sundsten et al, 2008b), goodness of fit tests have also been applied to mass spectrometry of blood Zhu et al, 2006).…”

Breast cancer detection using MRF-based probable texture feature and decision-level fusion-based classification using HMM on thermography images

“…forest and also other methods that perform the fusion in the feature or decision level [40][41][42][43]. But, they have the differences together in details and application.…”

“…While, in our model, all features intervene to decision and the classification is performed in a parallel state and simultaneously. While, the classification based on random forest and decision tree is carried out hierarchically and the features are used to decide in different steps of the tree [42][43].…”

Clinical proteomics in breast cancer: a review

“…For this competition, sera of breast cancer patients (n = 76) and healthy controls (n = 77) were analysed by MALDI-TOF MS. Data were subsequently analysed by ten competition participants for construction of diagnostic classifiers [52][53][54][55][56][57][58][59][60][61]. Surprisingly, though the various bioinformatic methods applied resulted in highly divergent classification models, reported performances (ranging from 83% to 89%) were very similar.…”