Mario Capecchi Laboratory, Uof U
Charles Keller, CCRI, UTHSCSA

The field of biological imaging is exploding, stemming, in part, from recent developments in imaging. This expansion also results from new scientific and biological applications for such data. These new applications do not necessarily adhere to the conventional clinical paradigm, in which a highly-trained radiologist looks at individual patients and makes specific diagnoses. Instead, biological imaging deals with populations of subjects, and the goal is typically not diagnosis but rather the evaluation of a particular hypothesis through the quantitative evaluation of images.

Image Based Phenotyping using the Mouse Hox Genes as a Prototype System

The Hox genes encode a group of transcription factors essential for proper development of the mouse. The expansion of Abdominal-B type Hox genes from a single Abd-B gene in invertebrates to a total of 16 genes in mammals suggests that they have played a significant role in the elaboration of more complex vertebrate morphologies. Mice with mutations in these genes manifest defects in the limb, vertebral column, urogenital system, and the caudal digestive tract. There appears to be a significant degree of functional redundancy among these genes as single mutants often have mild mutant phenotypes, while double and triple mutants are very severely affected. In previous studies, the Capecchi lab has shown that extra copies of Hoxd11 cause malformations in both forelimbs and vertebrae in mice. These studies were accomplished using physical isolation of the skeleton and microscopy for hand measurement of the forelimb bones. The Center is developing a workflow using a combination of CT imaging and image processing to automate much of the procedure of the original study. The combination of non-invasive imaging techniques and advanced image processing algorithms will allow for a substantial reduction in the time required per subject, thus allowing a greater number of studies and increased statistical power per study.

The analysis of skeletal geometry using a combination of CT imaging, image segmentation, and volume visualization was compared to a similar analysis using the combination of dissection, microscopy, and photo-imaging. We found that CT imaging combined with image segmentation offered a similar precision to that of the combination of dissection and microscopy with a significantly reduced time required. This opens the door to the possibly of statistically based high throughput studies of genetically modified mice using CT-imaging.