Category: Assay Development and Screening
New technologies for more efficient and less resource demanding plant breeding and plant cultivation are needed. Automation and high-throughput/high-content screening studies in single plant cells are largely lagging behind biomedical research on animal cells due to numerous reasons of technical and biological origin. Here we present an automated platform for quantitative phenotyping and perturbation studies of developing single plant cells, such as microspores and protoplasts. Relatively uniform populations of microspores, precursors of pollen cells, and protoplasts, plant cells lacking the cell wall, can be easily prepared from the floral buds or various plant organs, respectively (Figure 1A-C). Under optimal conditions, these cells can switch their cellular programs towards totipotency (Figure 1D), thus enabling formation of new embryo- or organogenic structures. Although this process is of high relevance in plant breeding and green biotechnology (e.g. Doubled Haploid technology) as well as in generation of novel and improved plants by using various gene technologies, due to poor understanding of key aspects beyond the reprogramming its utilization is highly limited and remains not applicable in many important crops and research plant models. Following mechanical or mechano-enzymatic isolation procedures, isolated cells can be cultured in vitro. We developed an efficient, multiwell-friendly cell immobilization procedure, which is optimal for automation of both, cell handling and cell analysis1 (Figure 2A-B) at various spatio-temporal scales (from sub-µm to mm and from minutes to weeks, Figure 2C). MORE multimode microscopy units for microscopy assays and a multimode plate reader for bioluminescence assays were integrated together with a liquid-handling unit, cell incubators and a robotic arm to establish a fully automated technology platform (Figure 3). Automation of miniaturized cultures enables a broad range of assays including pharmacogenetic perturbation studies driven by artificial intelligence and state-of-the-art computational approaches. Depending on assay type, throughput capacities may differ from dozens to hundreds of microtiter plates per day (Figure 4A). To handle and process the data, flow of which typically varies from multi-GB to TB-range per assay, expandable computing infrastructure was established and advanced computational approaches for image-driven analysis and efficient assay design were implemented. Our Q-DOT-A software application consist of two-modules for high-throughput image data processing and statistical data analysis, respectively, and enables easy configuration particularly for cytometry studies in various plant species (Figure 4B). Machine learning approaches, such as deep convolutional networks2 and active learning, are tailored for plant cell phenotyping applications, which require analysis of complex features (Figure 4C-D). Our automated platform provides a basis to improve plant cell technologies by linking quantitative phenotyping, perturbation studies and modelling3 in plant species- and in a customer-oriented manner. By further expanding to research on identification of physiological, molecular and genetic factors4 next generation technology in plant breeding and biotechnology is being established.
1 Dovzhenko et al.,2010,EP2455454A1
2 Ronneberger et al.,2015,LNCS,9351, 234-241
3 Johnson et al.,2017,Cytometry A 91,326-335
4 Chuprov-Netochin et al.,2016,BMC Plant Biology,16,192
Heiko Oehme– Research & Technology Development, Analytik Jena AG, Jena, Germany, Jena, Thuringen, Germany