Computational anatomy (CA) approaches start by estimating a canonical ‘template’ from a sample of images. This template is then used as a basis for statistical analysis to quantify structural differences among groups of interest. Here we apply CA to a sample of zebrafish with mutations in bmp1a and plod2 genes, which are associated with human brittle bone disease, and their unaffected siblings. Due to the complexity of fish skulls, previous attempts to classify craniofacial phenotypes have relied on qualitative features or 2D landmarks. In this work we aim to quantify 3D craniofacial phenotypes of zebrafish by comparing mutants to their wildtype siblings. We first estimate a ‘normative’ zebrafish template using microCT scans of the unaffected littermates as the sample pool using the Advanced Normalization Tools (ANTs). To validate the accuracy of the template and our CA pipeline, we compared the otolith volumes from the template CA approach to manually segmented volumes of the same set of zebrafish. Our CA based segmentation volumes are statistically indistinguishable from our manual segmentations and show that both mutants have larger otoliths than their wildtype controls. We are currently in the process of using the canonical template as a reference to conduct fully automated 3D shape analysis of our samples. Preliminary results suggest that phenotypic differences in both mutants are concentrated in the posterior portion of the frontal bone and in the dentary. We expect these methods will greatly improve the 3D analysis of the complex fish craniofacial phenotypes, especially those of zebrafish which are an important model system for testing genome to phenome relationships in the study of development, evolution, and human diseases.