Mouse Phenotyping by Bioimage Informatics
Mammalian genetics laboratory of NIG is studying relationship between genotype and phenotype, especially which gene causes which defect in the mouse body. They have technology to give birth to mouse with desired genes, so they compare the body of the mouse with/without a specific gene, and check the difference. For this purpose, they use X-ray CT scan images with a special contrast agent to see the internal body of the mouse with very high resolution, and now seeking a high-throughput image processging / computer vision techniques to identify the difference due to gene difference. Required techniques are 3D registration, non-rigid matching, shape analysis, and so on. This project is following the same direction with International Mouse Phenotyping Consortium.
2. Research Topics
The project is challenging the following research topics.
- Image registration - multiple micro-CT images are aligned so that we can make a mean image, or make a subtraction of two images.
- Detection of phenotypes - a micro-CT image is subtracted from a mean image so that we can make a list of candidates for phenotypes.
- Classification of phenotypes - some of the candidates correspond to known difference such as sex or blood clot, so we need to remove them to narrow down the list of candidates.
- Analysis of phenotypes - phenotypes are analyzed quantitatively. Sometimes we need a dedicated algorithm to analyze the part of the body in question.
- Image annotation - an annotation tool is developed for three dimensional micro-CT images so that we can use it for the evaluation of experiments.
More detail can be found in the following presentations.
- Project meeting, 2013-07-24, presentation slide
- Sharmili Roy, Xi Liang, Asanobu KITAMOTO, Masaru Tamura, Toshihiko Shiroishi, Michael S. Brown, "Phenotype Detection in Morphological Mutant Mice using Deformation Features", 16th International Conference on Medical Image Computing and Computer Assisted Intervention (MICCAI2013), 2013-09-25, poster
Some of our research results are released as tools.
- Sharmili Roy, Xi Liang, Asanobu KITAMOTO, Masaru Tamura, Toshihiko Shiroishi, Michael S. Brown,
"Phenotype Detection in Morphological Mutant Mice using Deformation Features",
16th International Conference on Medical Image Computing and Computer Assisted Intervention (MICCAI2013), Part III, LNCS 8151,
K. Mori et.al. (Eds.),