Input data type supported

• Preprocessed MALDI or LC-MS data.   
• Feature data – any CSV files containing intensity (or similar) values for consecutive samples and features, e.g. dataset obtained from any protein quantitation software or any ‘omics’ data such as gene expression microarray data.

Available algorithms

• One-dimensional data exploration
  • Measures of central tendency: mean, median, trimmed mean, minimum and maximum.
  • Measures of variability: standard deviation (std) and variance, sample range, interquartile range (IQR), median absolute deviation (MAD).
  • Empirical quantiles: the first and the third quartile (Q1, Q3).
  • Classical statistical tests: Wilcoxon-Mann-Whitney (WMW) test, Student’s t-test, Kolmogorov-Smirnov (KS) test.
  • Graphical tools: box plots, histograms.
• Multi-dimensional data exploration
  • Correlation between samples or between features.
  • Data visualization using heat maps.
• Dimension reduction
  
  • Feature extraction algorithms: PCA (Principal Component Analysis), LDA (Linear Discriminant Analysis), PLS (Partial Least Squares).
  • PCA visualization: Scree Plot, Scatterplot, Score & Loadings Scatterplot.
  • Advanced analysis techniques: Partial Least Squares combined with Logistic Regression (PLS-LR).
• Single biomarker detection
  
  • Separation measures available: Divergence, Fisher score, SAM score, T score, p-value of standard t-test, Kolmogorov-Smirnov test statistics, Wilcoxon-Mann-Whitney test statistics, Area Under ROC Curve (AUC).
  • Graphical tools: ranking chart, interactive Feature Browser.
• Multiple biomarker detection
  • Feature scoring algorithm: standard and normalized.
  • Performance criteria available: prediction accuracy, Matthews correlation.
  • Preliminary feature filters available: AUC (Area Under ROC Curve), Student’s t-test, Kolmogorov-Smirnov test, Wilcoxon-Mann-Whitney test, Divergence, Fisher score, SAM score, T score.
  • Classification algorithms available: LDA (Linear Discriminant Analysis), k-NN (k-Nearest Neighbors classifier), Decision Tree, SVM (Support Vector Machines), PLS (Partial Least Squares),
    Logistic Regression, randomForest — random tree ensemble.
  • Graphical tools: Feature Scores Bar Plot, Model Performance Line Plot, Heat Map, interactive Feature Browser.
• Classification models
  • Classification algorithms: k-NN — k-Nearest Neighbours algorithm, LDA – Linear Discriminant Analysis, SVM — Support Vector Machines, Tree — Classification Tree, PLS — Partial Least Squares, LR — Logistic Regression, randomForest — random tree committee.
  • Validation schemes: simple learning/test split, k-fold cross-validation, leave-one-out, random subsampling.
  • Feature extraction algorithms: PCA, PLS, LDA.
  • Feature selection algorithms: filter-based feature selection, stepwise feature selection methods (Forward/Backward Feature Selection).
  • Performance measures: misclassification error, specificity, sensitivity, AUC, permutation test, ROC analysis.
• Cluster analysis
  • Clustering algorithms: k-means, agglomerative clustering.
  • Validation tools: silhouette index, Davies-Bouldin index, Dunn’s
    index, C index.
  • Graphical tools: dendrogram, Validation Index Plot, 2D scatterplot.

Reporting/Exporting facilities

• Reporting of results with various report layouts for different formats (RTF, HTML, HTML presentation)
• Exporting of results for purposes of further analysis or for publication, including tables, charts, etc.