Semi-automatic classification of skeletal morphology in genetically altered mice using flat-panel volume computed...

Area of Science:

• Genomics and Bioinformatics
• Medical Imaging
• Computational Biology

Background:

• Mouse models are crucial for studying gene function, but rapid, noninvasive phenotyping methods are lacking.
• Identifying skeletal alterations in these models requires efficient screening tools.

Purpose of the Study:

• To develop and validate a novel, rapid, noninvasive screening method for detecting skeletal phenotypes in mouse models.
• To utilize computational intelligence for high-throughput analysis of mouse skeletal morphology.

Main Methods:

• Employed flat-panel volume computed tomography (fpVCT) for high-resolution 3-D imaging of mouse skulls.
• Developed a computational intelligence system, including a trained artificial neural network, for phenotype classification.
• Validated the method using discoidin domain receptor 2 (DDR2)-deficient mice with known skull abnormalities.

Main Results:

• fpVCT enabled high-contrast imaging with significant shape feature computation and visualization of morphological differences.
• The artificial neural network achieved highly accurate, semi-automatic classification of DDR2-deficient and wild-type mice.
• The method successfully identified subtle phenotypes in heterozygous DDR2 mice and classified DDR1 knockout mice.

Conclusions:

• The combination of fpVCT and artificial neural networks offers a reliable, cost-effective, and noninvasive primary screening tool for skeletal phenotypes in mice.
• This approach can identify novel mouse phenotypes with skull changes from various genetic models.
• Further development requires creating and training new neural networks for diverse applications.