Skip to content

Introduction

Hierarchically clustered matrices commonly represent high-dimensional biological data and are widely used for visualization. Although methods exist to assess cluster stability, dendrogram-defined clusters are rarely used for statistical inference. HiMaLAYAS is a general framework for post hoc enrichment-based annotation and visualization of hierarchically clustered matrices. It treats dendrogram-defined clusters as statistical units, tests categorical annotations for enrichment, controls multiple testing, and renders significant annotations alongside clusters.

Figure 1 HiMaLAYAS workflow and application to a hierarchically clustered yeast genetic interaction profile similarity matrix (Costanzo et al., 2016). A real-valued matrix and categorical annotations serve as inputs. HiMaLAYAS cuts the dendrogram at a user-defined threshold, tests categorical annotations for enrichment, controls multiple testing, and renders significant annotations alongside clusters.

Core Ideas

  • Hierarchical clustering organizes rows and columns into contiguous regions that can be treated as clusters.
  • Clusters are produced by cutting the dendrogram.
  • Enrichment is tested across dendrogram-defined clusters and categorical annotations.
  • Significant annotations are rendered alongside clusters for interpretation.
  • The workflow is domain-agnostic as long as you have a matrix and categorical annotations.
  • Enrichment can be evaluated at different dendrogram cut thresholds.

Typical Use Cases

  • Saccharomyces cerevisiae genetic interaction matrices (Costanzo et al., 2016) with GO Biological Process enrichment.
  • Expression similarity matrices with pathway enrichment.
  • Non-biological matrices such as country-sector input-output matrices with metadata-based enrichment labels.

The Basic Object Graph

  • Matrix: labeled numeric matrix used for clustering and plotting.
  • Annotations: term-to-label mapping aligned to the matrix label background.
  • Analysis: orchestrates clustering, enrichment, and layout.
  • Results: stores enrichment output and context for plotting.
  • Plotter: builds layered figures.