Email updates

Keep up to date with the latest news and content from BMC Bioinformatics and BioMed Central.

Open Access Methodology article

Detection of nuclei in 4D Nomarski DIC microscope images of early Caenorhabditis elegans embryos using local image entropy and object tracking

Shugo Hamahashi123, Shuichi Onami123* and Hiroaki Kitano124

Author Affiliations

1 Kitano Symbiotic Systems Project, ERATO, Japan Science and Technology Corporation, M31 6A, 6-31-15 Jingumae, Shibuya, Tokyo 150-0001, Japan

2 Graduate School of Science and Technology, Keio University, 3-14-1 Hiyoshi, Kohoku, Yokohama 223-8522, Japan

3 Institute for Bioinformatics Research and Development (BIRD), Japan Science and Technology Agency, 5-3 Yonbancho, Chiyoda, Tokyo 102-0081, Japan

4 Sony Computer Science Laboratories, Inc., 3-14-13 Higashi-Gotanda, Shinagawa, Tokyo 141-0022, Japan

For all author emails, please log on.

BMC Bioinformatics 2005, 6:125  doi:10.1186/1471-2105-6-125

Published: 24 May 2005

Abstract

Background

The ability to detect nuclei in embryos is essential for studying the development of multicellular organisms. A system of automated nuclear detection has already been tested on a set of four-dimensional (4D) Nomarski differential interference contrast (DIC) microscope images of Caenorhabditis elegans embryos. However, the system needed laborious hand-tuning of its parameters every time a new image set was used. It could not detect nuclei in the process of cell division, and could detect nuclei only from the two- to eight-cell stages.

Results

We developed a system that automates the detection of nuclei in a set of 4D DIC microscope images of C. elegans embryos. Local image entropy is used to produce regions of the images that have the image texture of the nucleus. From these regions, those that actually detect nuclei are manually selected at the first and last time points of the image set, and an object-tracking algorithm then selects regions that detect nuclei in between the first and last time points. The use of local image entropy makes the system applicable to multiple image sets without the need to change its parameter values. The use of an object-tracking algorithm enables the system to detect nuclei in the process of cell division. The system detected nuclei with high sensitivity and specificity from the one- to 24-cell stages.

Conclusion

A combination of local image entropy and an object-tracking algorithm enabled highly objective and productive detection of nuclei in a set of 4D DIC microscope images of C. elegans embryos. The system will facilitate genomic and computational analyses of C. elegans embryos.