A data mining tool for discovering subgroups of patients and genes that simultaneously display unusual levels of variability compared to other genes and patients. Based on sparse singular value decomposition (SSVD), the method can detect both mean and variance biclusters in the presence of heterogeneous residual variance.
