A fully automated deep learning pipeline for high-throughput colony segmentation and classification

Abstract

Background: Adenine auxotrophy is a commonly used non-selective genetic marker in yeast research.It allows investigators to easily visualize and quantify various genetic and epigenetic events by simply reading out colony color. However, manual counting of large numbers of colonies is extremely timeconsuming, difficult to reproduce and possibly inaccurate.Results: Using cutting-edge neural networks, we have developed a fully automated pipeline for colony segmentation and classification, which speeds up white/red colony quantification 100-fold over manual counting by an experienced researcher. Conclusions: Our approach uses readily available training data and can be smoothly integrated intoexisting protocols, vastly speeding up screening assays and increasing the statistical power of experiments that employ adenine auxotrophy. KeywordsDeep learning, neural networks, adenine auxotrophy, yeast, growth assay BackgroundAuxotrophy is the inability of an organism to synthesize a particular organic compound required for its growth. For example, yeasts with mutations in the adenine biosynthetic pathway cannot grow on media lacking adenine. When grown under adenine-limiting conditions, adenine auxotrophs grow but accumulate a cell-limited red pigment in their vacuoles, whereas wild type cells grow white. This red/white colony color assay has been widely used over the last decades for the investigation of many biological processes, such as recombination, copy number, chromosome loss, plasmid stability, prion propagation, or epigenetic gene regulation in both budding and fission yeasts. However, adapting this assay to quantitative high-throughput applications has proven challenging, as it requires extensive scoring of colony phenotypes by eye. In addition to being time-consuming and tedious, manual colony scoring may suffer from inaccuracy and irreproducibility. Nonetheless, up to now manual scoring is a common practice in the yeast community. Modern machine-learning techniques such as deep learning have made huge strides in automated image classification in recent years and are beginning to be applied to previously intractable problems in the biomedical imaging domain. We set out to leverage these recent developments to build a computational pipeline that would enable fully automated high-throughput adenine auxotrophy-based screening and quantification.Typically, red/white colony color assays start with plating individual yeast cells on limiting adenine indicator plates and allowing them to grow until they form colonies large enough to be inspected by eye.Each plate may represent an independent condition or genotype, the penetrance of which can be assessed by quantifying the percentage of non-white colonies per plate. In order to create a pipeline that would fit into existing protocols as seamlessly as possible, we considered our input to be images of plates and our desired output to be percentages of white and non-white colonies per plate. Two major tasks are required to generate this output: separating the colonies from th...

petiteFinder: An automated computer vision tool to compute Petite colony frequencies in baker’s yeast

“…It is also possible to treat cells directly with tetrazolium to induce a rapid transition from white to red, specifically in Grande cells (Hess et al, 2009). While the pipeline in (Carl et al, 2020) performs admirably, and could be modified for Petite detection with the red/white colony assay, there are two disadvantages to using the red/white assay over our experimental approach. First, the red/white assay requires experimental constraints on media (adenine limited media) and the genetic background of the strains, whereas large/small colony detection only requires constraints on the media.…”

“…Second, for red pigments to form, experimentalists often need to wait further after colony growth or apply chemical treatments which decrease experimental efficiency compared to our approach. With respect to the object detection pipeline in (Carl et al, 2020), one downside is that it requires post-processing of segmentation results, including heuristics on eccentricity and absolute size of colonies in pixels. Furthermore, assuming perfect segmentation in the study by Carl et al (where segmentation accuracy is not reported), petiteFinder still achieves higher accuracy in Grande/Petite classification compared to white/red classification in experiments.…”

“…For this reason we believe an exhaustive search of modifications to existing semi-supervised pipelines is unwarranted. There is, however, a supervised red/white colony detection pipeline described in (Carl et al, 2020) that can be used for Petite frequency calculations in strains with an appropriate genetic background. In yeast strains with specific mutations in the adenine synthesis pathway, an intermediate compound (initially white) accumulates and turns red when oxidized in respiring cells.…”

Plating and imaging of yeast colonies for segmentation and classification with neural networks

“…Further, if components of the adenine synthesis pathway are expressed at low level (pink phenotype), classification can vary between researchers and screens, potentially leading to erroneous and not fully reproducible results. To reduce these issues, we have developed a highly accurate and automated colony classification method based on neuronal networks with whole plate images as input (Carl et al;bioRxiv, doi: https://doi.org/10.1101/801845). In this protocol we describe in detail the procedures for plating and imaging as used for the pipeline training, which are crucial to achieve optimal results.…”