Background: Generation and phenotyping of mutant mouse models continues to increase along with the search for the most efficient phenotyping tests. Here we asked if a combination of different locomotor tests is necessary for comprehensive locomotor phenotyping, or if a large data set from an automated gait analysis with the CatWalk system would suffice.
New method: First we endeavored to meaningfully reduce the large CatWalk data set by Principal Component Analysis (PCA) to decide on the most relevant parameters. We analyzed the influence of sex, body weight, genetic background and age. Then a combination of different locomotor tests was analyzed to investigate the possibility of redundancy between tests.
Result: The extracted 10 components describe 80% of the total variance in the CatWalk, characterizing different aspects of gait. With these, effects of CatWalk version, sex, body weight, age and genetic background were detected. In addition, the PCA on a combination of locomotor tests suggests that these are independent without significant redundancy in their locomotor measures.
Comparison with existing methods: The PCA has permitted the refinement of the highly dimensional CatWalk (and other tests) data set for the extraction of individual component scores and subsequent analysis.
Conclusion: The outcome of the PCA suggests the possibility to focus on measures of the front and hind paws, and one measure of coordination in future experiments to detect phenotypic differences. Furthermore, although the CatWalk is sensitive for detecting locomotor phenotypes pertaining to gait, it is necessary to include other tests for comprehensive locomotor phenotyping.
Keywords: Activity; CatWalk; Locomotion; Mouse; Phenotyping; Principal component analysis.
Copyright © 2017 Elsevier B.V. All rights reserved.