loading ... please be patient ...
How do your datasets relate?
Choose a module by clicking its line. Each module integrates your data at the levels where its line stops: fusing the data directly (low), finding structure shared across layers (middle), or combining results (high).
Prepare filter · scale · QC Low level direct data fusion Middle level structure shared across layers High level combined results, pathways Vertical integration: different omics layers measured on the same (or partially overlapping) samples. Input: 2+ omics tables plus metadata, or feature lists. Vertical different omics, same samples K-means on joined layers Factors: MOFA, MCIA, DIABLO, NMF clusters (SNF) · cross-layer networks Pathways combined across layers · knowledge networks Start › Multi-platform fusion: several analytical platforms (NMR, GC-MS, LC-MS, Raman) on the same samples, merged into one block-weighted layer (low-level fusion) that continues as a vertical layer. Input: 2+ metabolomics tables plus metadata. Multi-platform fusion NMR, GC-MS, LC-MS, Raman; same samples merged into one layer, each platform block-weighted Start › Horizontal integration (meta-analysis): the same omics type measured in several studies on different samples. Input: 2+ tables of the same omics type, each with its own metadata. Horizontal same omics, different studies per-study QC Merged samples, pooled limma Combined p-values or effect sizes · pathway enrichment Start › Classification & Biomarkers: predict a phenotype and find the features that carry it; single layers, joined layers, DIABLO and late fusion compared under cross-validation. Input: 2+ omics tables on shared samples, plus metadata with group labels (50+ samples recommended). Classification & Biomarkers outcome to predict Classifiers on joined layers DIABLO multi-block Late fusion of per-layer predictions Start ›
Hover a line for what the module needs as input.