-
Categories
-
Pharmaceutical Intermediates
-
Active Pharmaceutical Ingredients
-
Food Additives
- Industrial Coatings
- Agrochemicals
- Dyes and Pigments
- Surfactant
- Flavors and Fragrances
- Chemical Reagents
- Catalyst and Auxiliary
- Natural Products
- Inorganic Chemistry
-
Organic Chemistry
-
Biochemical Engineering
- Analytical Chemistry
-
Cosmetic Ingredient
- Water Treatment Chemical
-
Pharmaceutical Intermediates
Promotion
ECHEMI Mall
Wholesale
Weekly Price
Exhibition
News
-
Trade Service
| It doesn’t take years and only a few hours |
| Machine learning can quickly reveal the internal structure of cells |
Science and Technology Daily, Beijing, October 10 (Reporter Zhang Mengran) Using high-power microscopes and machine learning, American scientists have developed a new algorithm that can automatically identify about 30 different types of organelles in ultra-high-resolution images of the entire cell And other structures
.
Related papers were published in the latest issue of "Nature" magazine
Aubrey Weigel, who leads the COSEM (cell segmentation under the electron microscope) project team, said that the details in these images are almost impossible to resolve manually in the entire cell
.
The data of just one cell consists of tens of thousands of images, and it takes more than 60 years for a person to track all the organelles of the cell through these images
In addition to the two articles in Nature, the research team also released a data portal "Open Organelles" through which anyone can access the data sets and tools they created
.
These resources are invaluable for studying how organelles keep cells running.
In the past decade, the research team has used high-power electron microscopes to collect a large amount of data from a variety of cells, including mammalian cells
.
The latest machine learning tools can pinpoint synapses, the connections between neurons, in electron microscope data
.
The researchers adjusted the algorithm to draw or segment the organelles in the cell.
The researchers said that using these numbers, the algorithm can also determine whether a particular combination of numbers is reasonable
.
For example, a pixel cannot be located both in the endoplasmic reticulum and in the mitochondria at the same time
To answer questions such as how many mitochondria are in a cell or what their surface area is, the research team built an algorithm that incorporates prior knowledge about the characteristics of organelles
.
After two years of work, the COSEM research team finally found a set of algorithms that can generate good results for the data collected so far
At present, the research team is improving imaging to a higher level of detail, and further optimizing tools and resources, creating a more extensive cell annotation database and more detailed images of cells and tissues
.
These results will support a new research field in the future-4D cell physiology, to understand the interaction of cells in the different tissues that make up an organism
