Automated control

Researchers from Tokyo Medical and Dental University (TMDU) have created a DeepACT artificial intelligence system that can identify healthy and productive skin stem cells more accurately and faster than a person examining each cell under a microscope.

Keratinocyte stem cells are used to treat hereditary skin diseases and to grow skin implants, which are used to repair burn injuries. This is one of the few types of adult stem cells that are easy to grow in the laboratory. Healthy keratinocytes move faster than less healthy cells, so they can be identified by the eye using a microscope. However, this process is time-consuming, time-consuming and error-prone.

Using a deep learning process, using a library of sample images, the researchers created the DeepACT artificial intelligence system, which automatically identifies and tracks the mobility and fitness of stem cells.

They tested DeepACT on a group of images and found that the results were very accurate compared to manual analysis.

DeepACT also calculates the movement index of each colony – a coefficient showing how fast the cells in the central part of the colony are moving compared to the cells in the peripheral region. The higher the movement index, the more likely this colony is to grow than with a lower movement index. Thus, this indicator allows us to determine the stem cells suitable for growing new skin needed by burn patients.

DeepACT includes two main modules: identification of human keratinocytes with single-cell resolution from phase-contrast images of cultures and tracking the movement of keratinocytes in the colony using a state space model. Since human keratinocyte stem cell colonies exhibit a unique pattern of movement, DeepACT can distinguish keratinocyte stem cell colonies from non-stem cell colonies by analyzing the spatial and velocity information of cells.

Skin grafting may be unsuccessful if the graft contains many weak or unproductive stem cells, and the ability to quickly and accurately identify the most suitable cells is an important clinical advantage. Automated quality control can also be useful for industrial stem cell production and will reduce production costs.

Article by T.Hirose et al. Label-free quality control and identification of human keratinocyte stem cells by deep learning-based automated cell tracking is published in the journal Stem Cells.

Aminat Adzhieva, portal "Eternal Youth" http://vechnayamolodost.ru based on TMDU materials: AI spots healthy stem cells quickly and accurately.