Hierarchical Clustering:

Manipulating Yeast Environmental Stress Response Microarray Data


Microarray technology has been heralded as the new biological revolution.  On a single "chip" it is possible to determine the relative expression levels of every gene within an organism's genome.  But examining an individual gene's expression levels without comparison to other genes or other experimental conditions is akin to thousands of telephone directories that each contain only a single listing -- useless.  In order to infer useful biological information and determine the relationships between individual genes, a system of "clustering" was needed that could group similarly expressed genes into sub-groups, and, therefore, categorize genes appropriately.

Determining the biological relationships between individual genes can be like searching for a needle in a haystack.  Consider all you have in front of you are 6,000 different expression ratios, each representing a gene in an organism's genome.  How are you to find any type of biological relationships from such a mass of data?  Enter the mathematicians.  They developed an algorithm known as hierarchical clustering to find similar patterns in large databases.  Basically, hierarchical clustering compares expression data and identifies those genes that are similar.

Use the links at left to learn more about microarrays, to learn more about hierarchical clustering, or to perform your own clustering analysis.


E-mail the authors:

amhartman@davidson.edu, sojohnson@davidson.edu, jekawwass@davidson.edu

Go to Davidson College, Math Department, or Biology Department


This page was designed for an undergraduate course, Computational Biology, at Davidson College.