A novel method for fast and statistically verified morphological characterization of filamentous fungi.

Along with productivity and physiology, morphological growth behavior is the key parameter in bioprocess design for filamentous fungi. Despite complex interactions between fungal morphology, broth viscosity, mixing kinetics, transport characteristics and process productivity, morphology is still commonly tackled only by empirical trial-and-error techniques during strain selection and process development procedures. In fact, morphological growth characteristics are investigated by computational analysis of only a limited number of pre-selected microscopic images or via manual evaluation of images, which causes biased results and does not allow any automation or high-throughput quantification. To overcome the lack of tools for fast, reliable and quantitative morphological analysis, this work introduces a method enabling statistically verified quantification of fungal morphology in accordance with Quality by Design principles. The novel, high-throughput method presented here interlinks fully automated recording of microscopic images with a newly developed evaluation approach reducing the need for manual intervention to a minimum. Validity of results is ensured by concomitantly testing the acquired sample for representativeness by statistical inference via bootstrap analysis. The novel approach for statistical verification can be equally applied as control logic to automatically proceed with morphological analysis of a consecutive sample once user defined acceptance criteria are met. Hence, analysis time can be reduced to an absolute minimum. The quantitative potential of the developed methodology is demonstrated by characterizing the morphological growth behavior of two industrial Penicillium chrysogenum production strains in batch cultivation.