COMPARING EVOLVABILITIES: COMMON ERRORS SURROUNDING THE CALCULATION AND USE OF COEFFICIENTS OF ADDITIVE GENETIC VARIATION
Version of Record online: 6 FEB 2012
© 2012 The Author(s). Evolution© 2012 The Society for the Study of Evolution.
Volume 66, Issue 8, pages 2341–2349, August 2012
How to Cite
Garcia-Gonzalez, F., Simmons, L. W., Tomkins, J. L., Kotiaho, J. S. and Evans, J. P. (2012), COMPARING EVOLVABILITIES: COMMON ERRORS SURROUNDING THE CALCULATION AND USE OF COEFFICIENTS OF ADDITIVE GENETIC VARIATION. Evolution, 66: 2341–2349. doi: 10.1111/j.1558-5646.2011.01565.x
- Issue online: 26 JUL 2012
- Version of Record online: 6 FEB 2012
- Accepted manuscript online: 5 JAN 2012 02:10AM EST
- Received August 12, 2011 , Accepted December 20, 2011
- natural selection;
- quantitative genetics;
- sexual selection
In 1992, David Houle showed that measures of additive genetic variation standardized by the trait mean, CVA (the coefficient of additive genetic variation) and its square (IA), are suitable measures of evolvability. CVA has been used widely to compare patterns of genetic variation. However, the use of CVAs for comparative purposes relies critically on the correct calculation of this parameter. We reviewed a sample of quantitative genetic studies, focusing on sire models, and found that 45% of studies use incorrect methods for calculating CVA and that practices that render these coefficients meaningless are frequent. This may have important consequences for conclusions drawn from comparative studies. Our results are suggestive of a broader problem because miscalculation of the additive genetic variance from a sire model is prevalent among the studies sampled, implying that other important quantitative genetic parameters might also often be estimated incorrectly. We discuss the most prominent issues affecting the use of CVA and IA, including scale effects, data transformation, and the comparison of traits with different dimensions. Our aim is to increase awareness of the potential mistakes surrounding the calculation and use of evolvabilities, and to compile general guidelines for calculating, reporting, and interpreting these useful measures in future studies.