According to experts, a novel computer pipeline for three-dimensional imaging data analysis can quickly and reliably show biologists how a plant’s leaves respond to its environment and help them find plants that use water more effectively.
A group of Penn State biologists and computer scientists created a 3D imaging model to investigate the interactions between nearby cells and the tiny stomatal guard cells, which are essential for transpiration and photosynthesis in plants.
The researchers discovered that the guard cells reacted in unanticipated ways, and the model is more effective and precise than current techniques for studying cellular geometry and mechanics.
The study will aid biologists in conducting trials more successfully and identifying plants, such as significant agricultural crops, that are better suited to climate change.
Dolzodmaa Davaasuren, a doctorate candidate in Penn State’s College of Information Sciences and Technology, spearheaded the construction of the pipeline.
“Currently, professionals take five to eight hours to manually label just the guard cells in a single 3D image set,” she said. To analyze more images, “our team aimed to automate operations.”
The model plant Arabidopsis thaliana, also referred to as thale cress, was used by the researchers in the construction and testing of their pipeline.
They created 3D photos of the plant’s defense cells using a specialized confocal microscope. Stomatal pores are surrounded by guard cells, which control the amount of carbon dioxide and water vapor that can travel through the pores.
To observe how stomatal volume changed, the team took pictures before and after ablating, or poking holes in, nearby cells that were touching guard cells using a laser beam.
The scientists built their model, which they named 3D CellNet, on the 3D U-Net segmentation model and added an encoder that better stores spatial information.
Additionally, an attention module was included, instructing the model to pay attention to certain areas of the 3D image.
They instructed the module to concentrate on the tiny guard cells in this instance. Just five 3D photos with manual labels were employed by the researchers to train their model.
To measure the guard cells’ morphologies, further image processing stages were added to the pipeline.
The team discovered that their new pipeline accurately and more quickly identified photos and quantified cell volumes than skilled cell biologists.
Additionally, they discovered that 3D CellNet segmentation outperformed its foundational model and two additional 2D models. In the journal Patterns, they published their research findings.
According to James Wang, distinguished professor of information sciences and technology and study co-author, “From a computer science perspective, this is the first time we’re able to use a machine trained with a limited number of labeled examples to achieve highly accurate 3D image segmentation in such a demanding situation.”
“Even while medical imaging has comparable 3D problems, it doesn’t have to deal with the difficulty of peering deeper inside a sample where the imaging becomes murkier due to light scattering.
The amount of light scattering increases with depth, but the exact amount is unclear. It’s a technical issue that we must overcome, and our research is one of the essential first steps in doing so.”
The scientists discovered that the guard cells behaved differently than anticipated in their responses to external stimuli using this new pipeline to segment and assess cell volume following ablation.
By eliminating the cells that border the guard cells, the researchers reasoned that the guard cell volume would grow and the pores would begin to open. The researchers, however, saw no improvement.
However, they discovered that guard cell volume increased when they abated the adjacent cells at the top and bottom of each guard cell pair, which are supposed to impede stomatal complex extension.
This forced the guard cells apart when they expanded and caused the stomatal hole to open.
Charles Anderson, associate professor of biology and research co-author, stated, “The nearby cells place mechanical limitations on the guard cells, but they’re doing so in a way that was completely unanticipated and might be somewhat independent of the water condition of those neighboring cells.
” We’d like to look into this further to learn more about the bio-mechanical processes that enable plants to efficiently close their stomata in response to drought and maintain that closure.
One of the most thrilling aspects of the paper is that it really is a tour de force in terms of computer science, developing a new algorithm that outperforms existing algorithms for measuring the 3D volumes of cells, and that it immediately applies that advance to help answer the crucial biological question of how stomatal pores, which are responsible for photosynthesis and water transport in plants, actually function.
In order to solve issues with food security in the face of a growing global population and climate change, tools like 3D CellNet can aid biologists in better understanding how guard cells and stomata respond to outside stimuli, Anderson said.
Source:Penn State
