This is the home page of the iCluster project for visualisation and statistical differentiation of subcellular imaging.
iCluster is a complete integrated methodology for testing for difference in fluorescent protein subcellular
localisation imaging. iCluster combines components for automated statistics generation,
spatial organisation and layout of image sets by similarity to enable patterns of difference in image sets
to be seen and browsed, and statistical testing to give rigorous confirmation of differences between experiments.
A movie showing iCluster in use can be found here (25M).
- Automated generation of statistics for distinguishing subcellular localisations
- Spatial layout of images by similarity in 2 or 3 dimensions
- Fully interative image browsing: rotate around, zoom in, select, reclassify, create new classes, hide images
- Easy visual detection of outlier/unusual images
- Statistical testing for difference between image sets, i.e. comparing treated/untreated cells
- Automated representative image selection to find the image that best represents an experiment
- Can be applied to organising and browsing virtually any image set by supplying your own image quantifiers
- Image statistics may be individually selected/deselected for use in creating the visualisation
or statistical testing for difference
Statistical and visual differentiation of high throughput subcellular imaging, N. Hamilton, J. Wang, M.C. Kerr and R.D. Teasdale, BMC Bioinformatics 2009, 10:94. link.
Visualising and clustering high throughput sub-cellular localization imaging, N. Hamilton & R.D. Teasdale. BMC Bioinformatics 9:81, 2008. link.
Further details may be found in the user manual available on the download page.
iCluster is built with the very beautiful Processing.
(c) 2009 Nicholas Hamilton. All rights reserved.