Skip NavigationSkip to Content

Automatic Segmentation Software and Application to High-throughput Gene Localization in Interphase Nuclei

Prabhakar Gudla, Kaustav Nandy, Stephen Lockett, Dean McCullough, Tom Misteli, Karen Meaburn

High-throughput 2D segmentation of nuclei in cell culture samples

Automatic segmentation (delineation) of cell nuclei is essential for many high-throughput cytometry applications. Our software is unique because it adapts on the fly to changes between data sets in the characteristics of the imaged nuclei.

2-D image with DAPI stained nuclei. Scale bar at top right corner = 15 mm

When tested on 7 different data sets comprising 4 different lymphoblast cell lines and containing 4,181 nuclei where 47% were clustered based on visual inspection, 3,515 were detected as single nuclei (sensitivity: 84%), of which 3,507 were correctly segmented (specificity: 99.8%).

Segmentation Result. Red: single nuclei; blue: clustered nuclei


High-throughput 2D segmentation of nuclei in tissue sections (In Progress)

The genome is non-randomly organized within the cell nucleus. Misteli et al have identified several genes that are specifically repositioned during mammary tumorigenesis, are unrelated to gene activity, and distinguish different tumor types. To significantly expand the scope of these studies, we are in the process of automating the screening of tissue samples for gene positions.
Tissue section showing FISH signals from gene loci (red and green) and nuclei (blue)
Segmentation of the Cell Nuclei (preliminary result)


OMAL Image Analysis Toolbox

Page last updated November 3, 2010 @ 4:00 pm