MetaStat performs data preparation and statistical analysis on feature tables generated by MetAnalyzer or from externally supplied datasets. Data preparation features include missing-value imputation, blank subtraction, feature filtering, normalization (QC-based and global intensity methods), and transformation and scaling. Statistical analysis includes univariate methods (t-test, ANOVA, fold-change analysis) for hypothesis testing and multivariate approaches (PCA, PLS-DA, hierarchical clustering) for pattern discovery and dimensionality reduction, with integrated visualization. A number of feature selection methods are also implemented to identify discriminative variables based on statistical significance, effect size, or model-based importance metrics.
