SAI: Fast and automated quantification of stomatal parameters on microscope images

Abstract

Using microscopy to investigate stomatal behaviour is a common technique in plant physiology research. Manual inspection and measurement of stomatal features is a low throughput process in terms of time and human effort, which relies on expert knowledge to identify and measure stomata accurately. This process represents a significant bottleneck in research pipelines, adding significant researcher time to any project that requires it. To alleviate this, we introduce StomaAI (SAI): a reliable and user-friendly tool that measures stomata of the model plant Arabidopsis (dicot) and the crop plant barley (monocot grass) via the application of deep computer vision. We evaluated the reliability of predicted measurements: SAI is capable of producing measurements consistent with human experts and successfully reproduced conclusions of published datasets. Hence, SAI boosts the number of images that biologists can evaluate in a fraction of the time so is capable of obtaining more accurate and representative results.

Cell type-specific responses to fungal infection in plants revealed by single-cell transcriptomics

“…For this purpose, confocal imaging was conducted on Arabidopsis leaf tissues inoculated with C. higginsianum to capture images of guard cells at 40 hpi. The width/length ratio was measured for each stomata using a machine learning-based imaging acquisition workflow (Sai et al, 2022) to quantify stomatal aperture. We measured the stomatal aperture in guard cells at the infection sites, in the more distal regions but still the neighbouring tissue of the invasive hyphae, and in mocktreated tissues.…”

“…At 40 hpi, the samples were collected for imaging acquisition. The image files capturing the stomatal aperture of uninfected, distant, or proximity guard cells were fed into a machine learning-based analysis pipeline (Sai et al, 2022). The width and length of each guard cell were measured and collected for comparison.…”

Rapid non-destructive method to phenotype stomatal traits

“…Recently, a number of methods have been published which address the issues with manual processing by including a machine learning algorithms that automatically detect and analyse stomatal features (16)(17)(18)(19)(20). While this improvement significantly accelerated image analysis and offers an effective substitute to manual analysis, these methods are limited by providing either information on stomatal number, size or aperture in one species in isolation; having data on all stomatal traits is desirable to have a comprehensive view on stomatal phenotype.…”

Image-based quantification ofArabidopsis thalianastomatal aperture from leaf images

“…Stomatal detection by Faster Region-based CNN (Faster R-CNN) and ChanVese-based pore segmentation was applied for poplars ( Populus L.) (Li et al, 2019). Mask R-CNN was implemented to both count stomata and quantify the stomatal aperture of black poplar ( Populus nigra ) and gingko ( Gingko biloba ) (Song et al, 2020) or epidermal peels of Arabidopsis and barley ( Hordeum vulgare ) (Sai et al, 2022).…”

“…In this study, we aimed to develop an image analysis module that can quantify stomatal aperture of Arabidopsis leaves to accelerate routine benchtop analysis involved in physiological research. Up-to-date methods to quantify Arabidopsis stomatal aperture from brightfield microscope images have been limited, with the exception of a previous study (Sai et al, 2022). Notably, the module built in the Sai et al study was specifically designed for epidermal peels.…”