Fitting an order (7) seasonality adjusted PAR model, we computed within sample predictions and found the mean-squared error (MSE) to be 3.48, which is marginally better than using predictions based on the estimated mean number of daily admissions (MSE = 3.55).
All negative binomial regression models, out-of-sample prediction and estimation of 95% PI were conducted in R (V.2.15.1).
The agreement between PAM50 centroid-based single sample prediction (SSP) and PLS-regression was excellent (weighted Kappa: 0.988) within the training samples, but deteriorated substantially in independent samples, which could attribute to much more unclassified samples by PLS-regression.
The squared correlation between out-of-sample predictions and observed values was 0.14.
The three main steps involved in evaluation are: (i) identification of a suitable metric (such as the area under the ROC and root mean square between prediction and measurement); (ii) simulation of a null distribution for the chosen metric by evaluation of randomly sampled predictions; and (iii) assignment of a P-value for a prediction with respect to the null distribution for the metric.
Unsupervised clustering and sample prediction analyses using genes reported as PR-A and PR-B-specific targets further showed that BRCA-mutated samples did not exhibit an altered PR response.
It integrates patients information and MS data storage, process, sample classification and sample prediction into a single, user-friendly workbench.
Using the CCNB1 gene and a Core Serum Response gene signature as two examples we illustrate the usefulness of GOBO for rapid online analysis in a large breast cancer data set The web interface of GOBO allows precompiled data sets to be queried by the three main applications of GOBO: Gene Set Analysis (GSA), Co-expressed Genes (CG), and Sample Prediction (SP) (Figure 1).
BioSunMS provides the Wizard for general users, including extraction of MS peaks, construction of m/z matrix, sample classification and sample prediction.
In contrast to FTEn versus FTEb, PR-dependent genes could successfully differentiate malignant versus non-malignant samples, utilising both unsupervised clustering and sample prediction analysis.
Model validity was assessed by consulting the associated goodness of fit of the model (Q), along with the measures of CV-ANOVA and sample prediction from within SIMCA (Umetrics, Umea, Sweden) [ 15].
