Preprocessing Module- Input data by using the wizard method and focused on simplicity of data filtering.
- Global normalization, Lowess normalization (arraywise/blockwise)
- Embodied Global normalization, Lowess normalization (arraywise/blockwise)
DEG Module- Embodied fold change and various kinds of statistic algorithms (Welch¡¯s t-test, one-way ANOVA etc)
- Contains the volcano plot that can compare the fold change and the results of statistic test intuitively.
- Able to understand the distribution of each sample group in three-dimension by using the sample PCA.
Clustering Module- Embodied main clustering algorithm (hierarchical clustering, K-means, SOM)
- Provide various linkage method of Hierarchical clustering
- Excellent visual abilities such as dendrogram, topographic 2D cluster profile etc.
- Applied statistical clustering validation method to find the most appropriate clustering method.
Classification Module- It systemized the whole process of classification into feature selection, classifier and error estimation parts,
and embodied the whole computation function that can perform the whole process at once.
- Embodied the widely used method in the feature selection (BSS/WSS, regularized t-test, Kruskal-Wallis test etc).
- Embodied the widely used methods in the classifier (weighted KNN, prototype matching, FLDA etc).
- Embodied the widely used method in error estimation (complete/incomplete LOOCV, bootstrap etc).