Discover Hidden Molecular Phenotypes
TopMD’s TDA BioMapper leverages Topological Data Analysis to reveal hidden structures in complex omics data — discovering molecular phenotypes that traditional clustering methods miss.
Builds intuitive network maps
- Nodes represent groups of molecularly similar samples
- Edges connect overlapping samples, showing continuity between states
- Automatically detects molecular phenotypes based on omics similarity
- Interactive exploration of disease heterogeneity, treatment response, and progression
- Outputs interactive HTML networks, phenotype assignments, and statistical summaries
Phenotypes rooted in topology — not assumptions
After constructing the network, BioMapper applies community detection to find groups of nodes that are more tightly connected internally than externally. These communities represent distinct molecular phenotypes, defined by real structural separation in the data.
Network-aware Pathway Activation
Measure activation of an entire pathway as a shape-informed signal. We compute a pathway’s integrated activity relative to the global biological interaction network — capturing coordinated behaviour across genes, not just individual changes. The result is a robust, mechanistically meaningful score you can use to stratify patients and link phenotypes to biology and drug response.
- Whole-pathway “volume” from shape in global omics
- Network-context aware and robust to noise
- Produces interpretable, mechanism-tied stratification