Efficient phenotyping is essential for the identification of relevant changes and phenotypes and for the reproducibility of scientific experiments. A number of automated tests have been developed to speed up data acquisition, to make the results more objective and replicable and to reduce the burden on the animal through less handling. In addition, the number of acquired parameters per test is drastically increased. However it is not clear which tests and parameters are the most valid, reliable and valuable.
The authors addressed this question and analysed tests for locomotor phenotypes, whereby the definition is used very broadly. Included are gait phenotypes, as analyzed by Catwalk (CW), as well as activity, (as in the Open Field, home cage and SHIRPA) and motor ability phenotypes (as in the Grip Strength, Rotarod and Vertical Pole test). Gait phenotypes describe how the animal moves, activity phenotypes describe how frequently the animal moves, and motor ability phenotypes describe muscle strength, balance and coordination. The authors wanted to know if the CW test alone, with its high number of parameters, can be sufficient to detect locomotor deficits and if the data set can be reduced in dimensions by Principal Component Analysis (PCA) to detect meaningful components to ease data handling.
The PCA revealed a set of ten principal components. These components describe 80% of the total variance in the CW data and characterize different aspects of gait. With these components it was possible to detect effects of the used CW version, sex, body weight, age and genetic background. The PCA on a combination of locomotor tests suggests that these are independent without significant redundancy in their locomotor measures, thus emphasizing their originality.
The authors conclude that the primary focus in CW analysis could be put on front paw and hind paw measures (not all individual paws) as well as one of the interlimb coordination measures thus reducing the number of parameters for analysis. Although the CW is sensitive to detect relevant gait phenotypes, it cannot account for all aspects of locomotor phenotypes.
Zimprich A, Östereicher MA, Becker L, Dirscherl P, Ernst L, Fuchs H, Gailus-Durner V, Garrett L, Giesert F, Glasl L, Hummel A, Rozman J, de Angelis MH, Vogt-Weisenhorn D, Wurst W, Hölter SM. Analysis of locomotor behavior in the German Mouse Clinic. J Neurosci Methods. 2017 May 5. pii: S0165-0270(17)30124-3. doi: 10.1016/j.jneumeth.2017.05.005. [Epub ahead of print]