Spectrum

Spectrum is a spectral clustering tool developed to integrate and cluster multi-omic data for identifying disease subtypes and reduce noise. The software tool uses a self-tuning density-aware kernel and tensor product graph data integration with a diffusion procedure to reveal underlying structures. Spectrum contains a new method for finding the optimal number of clusters involving eigenvector distribution analysis. The software tool provides improved runtimes and better clustering results compared to other methods across 21 real expression datasets.