Application of the Taguchi method in poultry science: estimation of the in vitro optimum intrinsic phytase activity of rye, wheat and barley

Abstract

1. In poultry investigations, the main interest is often to study the effects of many factors simultaneously. Two or three level factorial designs are the most commonly used for this type of investigation. However, it is often too costly to perform when number of factors increase. So a fractional factorial design, which is a subset or a fraction of a full factorial design, is an alternative. The Taguchi method has been proposed for simplifying and standardising fractional factorial designs.

2. An experiment was conducted to evaluate the applicability of the Taguchi method to optimise in vitro intrinsic phytase activity (IPA) of rye, wheat and barley under different culture conditions.

3. In order to have a solid base for judging the suitability of the Taguchi method, the results of the Taguchi method were compared with those of an experiment that was conducted as a 3⁴ full factorial arrangement with three feed ingredients (rye, wheat and barley), three temperatures (20°C, 38°C and 55°C), three pH values (3.0, 5.5 and 8.0) and three incubation times (30, 60 and 120 min), with two replicates per treatment.

4. After data collection, a Taguchi L 9 (3⁴) orthogonal array was used to estimate the effects of different factors on the IPA, based on a subset of only 9 instead of 81 treatments. The data were analysed with both Taguchi and full factorial methods and the main effects and the optimal combinations of these 4 factors were obtained for each method.

5. The results indicated that according to both the full factorial experimental design and the Taguchi method, the optimal culture conditions were obtained with the following combination: rye, pH = 3, temperature = 20 °C and time of incubation = 30 min. The comparison between the Taguchi and full factorial results showed that the Taguchi method is a sufficient and resource saving alternative to the full factorial design in poultry science.

ACKNOWLEDGEMENTS

This research was supported by the office of vice president in research at the Ferdowsi University of Mashhad, Iran (project no. 3/18166). The authors would like to thank the laboratory staff of the Animal Science Department, University of Wageningen, for their contribution.