Investigating the Use of Fractional Replication and Taguchi Techniques in Cryopreservation: A Case Study Using Orthodox Seeds of a Tropical Rainforest Tree Species

Abstract

Cryopreservation is increasingly important for conserving endangered species including tropical rain forest germplasm where optimal cryopreservation protocols must be established rapidly sacrificing little germplasm. Currently, full factorial experiments analysed by ANOVA establish optimal conditions. However, these experiments can contain many treatment combinations whilst ANOVA identifies significant effects without guaranteeing to find robust optimal conditions. Taguchi optimization techniques efficiently identify robust conditions through fractional factorial experiments with an appropriate signal to noise ratio (SNR). This paper reports for the first time the use of Taguchi techniques in cryopreservation. An orthodox seed (Cassia siamea Lam.) was used to guarantee sufficient data to compare full and fractionally replicated experiments analysed using both ANOVA and SNR. For sprouting day (smaller is better), identical significant main effects were found for all experimental sizes for ANOVA and SNR. The 1/4 replicate did not allow investigation of interaction terms, but the significant main effects were the same for larger experiments. For shoot to root ratio (nominal is best), a significant main effect was found for all experimental sizes using ANOVA. This was also found using SNR, which identified additional significant main effect and interactions. No significant effects were found for dry weight (larger is better). We show smaller experiments are possible, provided important two level interactions are analysed. Differences determining the optimal treatment combination were found between ANOVA and SNR; with the Taguchi choice provides more robust solutions. Taguchi optimization techniques are recommended when germplasm is scarce and/or the experiment needs to be conducted rapidly.