Sci Rep: Scientists have developed new algorithms that could prevent misrecognization of cancer cells!

December 17, 2020 // -- In a recent study published in the international journal Scientific Reports entitled "Towards image-base cell cell lines using deep neural networks", scientists from institutions such as the University of Kent have developed a new type of computer algorithm that may be able to identify differences in cancer cell lines based on microimaging techniques.
Cancer cell line is a special type of cell that is removed from the lab and can grow with cell cultures as a nutrient and can be used by researchers to study and develop new cancer drugs;
This may have been a persistent problem since scientists began studying cancer cell line, and short tandem repeat analysis is often used to identify cancer cell line, but it is expensive and time-consuming;
Photo Source: Pixabay/CC0 Public Domain In this study, scientists trained computer models by using comparative analysis of large amounts of cancer cell data over time based on microimaging from a large number of cell lineages and using computer models for deep learning.
, they developed a new algorithm that allows computers to analyze individual microscopic images of cell lineages and accurately identify and tag them.
breakthrough could provide researchers with an easy-to-use tool to quickly identify and analyze all cell linetes in the lab without professionals and equipment.
researchers say our results have proven great potential for implementation in laboratory and cancer research, and using this new algorithm will yield further results and hopefully change the way cells are identified in scientific research, and give scientists a better chance of accurately identifying cells to reduce the occurrence of errors in cancer research and potentially save patients' lives.
addition, the researchers note that the new computer model can also accurately assign criteria for correctly identifying cell line, meaning that future researchers' ability to train to accurately identify cells may be further enhanced.
original source: Mzurikwao, D., Khan, M.U., Samuel, O.W. et al. Towards image-based cancer cell lines authentication using deep neural networks. Sci Rep 10, 19857 (2020).doi:10.1038/s41598-020-76670-6。