Automated monitoring of pollen quality in Douglas fir using digital images: applications for handling and storage

Abstract

Background: Conservation and breeding programs for forest species require efficient pollen quality monitoring strategies to maximize collection, storage and pollination efforts.
 Question: Can the ImageJ image analysis software be implemented to efficiently monitor pollen viability and germinability for quality testing?
 Studied species: Pseudotsuga menziesii (Mirb.) Franco
 Methods: Tetrazolium staining and germination on 10 % Brewback media was used to test Douglas-fir pollen quality. Two macros for automated counting were developed with ImageJ free software using digitized microscopic images, and were tested to evaluate provenances, germination in different media, dehydration/rehydration and freeze-storage treatments, and finally to compare pollen quality between mature trees and GA-induced strobili from juvenile trees.
 Results: Automated monitoring of pollen quality was feasible because the software suitably distinguishes viable from non-viable pollen grains due to differential tetrazolium staining that translates into pixel density, whereas germinability was assayed based on pollen size and circularity. This tool was used to guide the selection of appropriate pollen handling conditions and compare pollen production in adult and GA-induced juvenile trees. Pollen flash-freezing and storage at -80 °C showed recovery between 74 and 98 % of its initial germinability after rehydration; however, viability detected by tetrazolium is reduced, suggesting that handling and storage affect the reliability of the test.
 Conclusions: The flexibility of the Fiji software allowed the development of an efficient and accurate macro for monitoring P. menziesii pollen quality, reducing invested time and effort involved, which supports future research on its conservation physiology and its successful application for germplasm production.