About this supplement

Installation

Widget Catalog

D. discoideum example

Schema
   File
   K-Means Clustering
   Expression Profiles
   Heat Map
   Scatterplot
   Expression Profiles (3)
   Heat Map (2)
   GO Term Finder
   Data Table
   Expression Profiles (2)
   Genome Map
   GO Term Finder (2)

S. cerevisiae example

Spotlight

Our widgets for functional genomics use Orange, a data mining and machine learning suite. Orange can be accessed through scripting in Python, or by visual programming in Orange Canvas.

In functional genomics, we have also designed a web-based tool for mutant data analysis called GenePath (also featured in Science's NetWatch).

FRI > Biolab > Supplements > Microarray Data Mining with Visual Programming > D. discoideum Example > K-Means Clustering widget

K-Means Clustering widget

First step in our analysis is to cluster genes based on their expression profiles. We have used the k-means clustering algorithm, with k set to 5 and the Euclidean measure for distance.

Input of this widget are data tokens coming from the "File" widget. The output of the "K-Means Clustering" are new data tokens where genes are classified into k=5 clusters. Statistics giving the number of genes (Items) in each cluster and two quantitative measures of cluster quality are displayed.