A novel computational device can systematically analyze cell images in C. elegans

A joint research team co-led by the City University of Hong Kong (CityU) has developed a state-of-the-art computing device that can recreate and visualize three-dimensional (3D) shapes and temporal changes of cells, speed up the monitoring process from hundreds of hand hours to a few hours with the computer.

Refreshing the way biologists analyze imaging data, this tool can advance further studies in developmental and cell biology, such as cancer cell growth.

The interdisciplinary study was co-led by Professor Yan Hong, Chair of Computer Engineering and Professor of Data Engineering Wong Chung Hong in the Department of Electrical Engineering (EE) at CityU, together with biologists from Hong Kong Baptist University (HKBU) ) and Peking University.

Their findings were published in the scientific journal Nature Communication, entitled “Establishing a morphological atlas of Caenorhabditis elegans embryo using 4D separation based on in-depth learning“.

The tool developed by the team is called “CShaper”. “It is a powerful computational tool that can systematically separate and analyze cell images at a single-cell level, which is much needed for the study of cell division, and cell and gene function,” said Dr. Yan account of.

The bottle in analyzing the bulk of cell division data

Biologists have studied how animals grow from a single cell, a fertilized egg, to organs and the body as a whole through countless cell divisions. In particular, they want to experience gene functions, such as the specific genes involved in cell divisions for the formation of different organs, or what causes the abnormal cell divisions that follow to tumor growth.

One way to find the answer is to use the gene knockout method. With all the genes present, researchers first obtain cell images and the tree line.

They then “pull down” (remove) a gene from the DNA sequence, and compare the two linear trees to analyze changes in the cells and detect gene functions . They are then re-tested with other genes.

In the study, the team used a collaborative biologist Caenorhabditis elegans (C. elegans) embryos to extract terabytes of data for Professor Yan’s team to perform computer analysis. C. elegans this is a type of worm that shares many essential biological features with humans and is a valuable model for studying the process of tumor growth in humans.

“With an estimated 20,000 inputs C. elegans, it means that nearly 20,000 tests would be needed if they lay down one gene at a time. And there would be a lot of data. It is therefore essential to use an automated image analysis system. And this makes us develop a more efficient one, “he said.

Breakdown in automatically separating cell images

Cell images are usually obtained by laser beam scanning. Existing image analysis systems can only detect the nucleus of a cell well by the image quality of a cell, inhibiting the reconstruction of cell shapes.

Also, there is a lack of a reliable algorithm for separating 3D images with time (i.e. 4D images) of cell division. Image rotation is a critical process in computer vision that involves dividing visual input into sections to simplify image analysis. But researchers have to spend hundreds of hours labeling many cell images by hand.

The break in CShaper is that it can detect organs, build cell shapes in 3D, and more importantly, automatically separate the cell images at the cellular level. “Using CShaper, biologists can determine the content of these images within a few hours.

It can identify cell shapes and surface structures, and provide 3D views of cells at different times, “said Cao Jianfeng, a PhD student in Professor Yan ‘s group, and co – author of the paper.

To achieve this, a key part of the CShaper system is the DMapNet-based in-depth learning module developed by the team.

“By learning to capture several different distances between image pixels, DMapNet captures the shape of the organs while considering shape information, rather than just intense features. that CShaper achieved 95.95% accuracy in identifying the cells, which outperformed the other methods, “he explained.

With CShaper, the team created a real-time 3D atlas of cell morphology for the C. elegans embryo from the 4- to 350-cell stages, including cell shape, volume, surface area, migration, nucleus positioning and cell-cell communication with a proven cell identity.

Further research into tumor growth

“As far as we know, CShaper is the first computer system to separate and analyze the images of it. C. elegans embryo routinely at the single-cell level, “Mr. Cao said. Through close collaboration with biologists, we proudly developed a useful computer tool for the automated analysis of large amounts of cellular data.

We believe it can stimulate further studies in developmental biology and cell biology, especially in understanding the origin and growth of cancer cells, ”Dr. Yan said.

They also tested CShaper on plant stem cells, showing promising results. They believe that the computer device can be adapted to other biological studies.