Identification and quantification of sources of biological variance: a methodological approach.
Ketelaere B. de, Stulens J., Lammertyn J., Baerdemaeker J. de
Author Affiliation: K.U.Leuven, Department of Agro-Engineering and -Economics, Laboratory for Agricultural Machinery and -Processing, Kasteelpark Arenberg 30, 3001 Leuven, Belgium.
: 523-529
Abstract : A correct identification and quantification of the different sources of variance in a recorded dataset is of utmost importance in many ways, for instance when comparing treatment groups, or in the case where there is a need for describing the (future) behaviour of a batch of biological products. The total data variance can be split up into 2 different parts, one describing the biological variance (due to the natural heterogeneity of the batch), and the other describing the uncertainty (due to the imperfect measurement of the attribute considered). The classical approach to include biological variance is to use a 2 stage approach in which in a first stage, a (nonlinear) model is built for each product individually, where after inferences are based on the parameters obtained from the first stage. In this contribution, a methodological approach was proposed to identify and quantify the different sources of biological variance, using the concept of (nonlinear) mixed effects models. Such models are specifically designed to handle repeated measures data with high biological variance. The concept is demonstrated using a practical dataset of postharvest firmness changes in mangoes. It is shown that aside from the differences in biological age of the mangoes as a variance source, also the decay rate varies among mangoes. Furthermore, it is shown that at harvest, the biological variance is the dominating source of variance, whereas near the end of the storage period, the uncertainty about the measurements dominates.