Articles
IDENTIFICATION AND QUANTIFICATION OF SOURCES OF BIOLOGICAL VARIANCE: A METHODOLOGICAL APPROACH
Article number
674_68
Pages
523 – 529
Language
English
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 two 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 two 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, we propose a methodological approach to identify and quantify the different sources of biological variance, using the concept of (nonlinear) mixed effects models.
Such models are specifically designed in order 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.
The total data variance can be split up into two 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 two 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, we propose a methodological approach to identify and quantify the different sources of biological variance, using the concept of (nonlinear) mixed effects models.
Such models are specifically designed in order 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.
Authors
B. De Ketelaere, J. Stulens, J. Lammertyn, J. De Baerdemaeker
Keywords
nonlinear mixed effects models, postharvest quality, modelling
Online Articles (81)