HAllA (Hierarchical All-against-All association) is a method for finding blocks of associated features in high-dimensional datasets measured from a common set of samples. HAllA operates by 1) optionally discretizing mixed continuous and categorical features to a uniform representation, 2) hierarchically clustering each dataset separately to generate a pair of data hierarchies, 3) performing all-against-all association testing between features across two datasets using robust measures of correlation, 4) determining the statistical significance of individual associations by permutation testing, and 5) iteratively subdividing the space of significant all-against-all correlations into blocks of densely associated occurring as clusters in the original datasets.
module spider halla to find out what environment modules are available for this application.
- HPC_HALLA_DIR - installation directory
- HPC_HALLA_BIN - executable directory
If you publish research that uses halla you have to cite it as follows:
Gholamali Rahnavard, Kathleen Sucipto, Lauren J. McIver, Emma Schwager,Jason Lloyd-Price, George Weingart, Yo Sup Moon, Xochitl C. Morgan, Levi Waldron, Eric A. Franzosa, Curtis Huttenhower. High-sensitivity pattern discovery in large multi'omic datasets. (huttenhower.sph.harvard.edu/halla)