PermutMatrix is a specialized, free Microsoft Windows graphical software used to discover patterns in complex datasets by combining hierarchical clustering analysis (HCA) and matrix seriation. Originally designed for biological gene expression profiles, it is widely utilized across archaeology, sociology, and psychology to reorganise the rows and columns of a data matrix into an optimal linear order.
By adjusting the visual layout of a heatmap, it helps you visually isolate hidden structures and relationships that standard text tables obscure. Core Concepts: Clustering vs. Seriation
To use the tool effectively, it is vital to understand how it merges two distinct data mining techniques:
Hierarchical Clustering: Groups similar items into a nested, tree-like structure called a dendrogram. While it determines which items belong together in a branch, it does not define a strict left-to-right order for those branches.
Seriation: Linearly arranges items along a continuous spectrum based on a mathematical dissimilarity matrix. The objective is to place the most similar items as close to each other as possible, minimizing overall “stress” or noise across the grid. Step-by-Step Workflow in PermutMatrix 1. Importing and Normalizing Data
Format: Prepare your numeric dataset in a standard text format (tab or comma-separated) or Eisen’s Cluster format.
Transformation: Use the software’s preprocessing options to normalize your data (e.g., log-ratio transformations or centering). This prevents high-magnitude rows from skewing the distances. 2. Selecting a Distance Metric
PermutMatrix builds a square symmetrical dissimilarity matrix (D) measuring the resemblance between pairs of elements. You must choose an analytical expression to calculate this distance: www.atgc-montpellier.fr
What is seriation – ATGC: Montpellier Bioinformatics platform
Leave a Reply