AI can identify single cells with disease
Human Cell Atlas is the largest and fastest-growing single-cell atlas in the world. The Human Cell Atlas contains millions of references to cells in tissues, organs, and developmental stages. These references allow physicians to better diagnose and treat their patients by understanding the effects of disease, aging and environment on cells. Reference atlases are not without their challenges. The single-cell datasets can contain measurement errors due to the batch effect, the availability of computational resources globally is limited and sharing raw data often is restricted by law.
Researchers at Helmholtz-Zentrum München and the Technical University of Munich have developed a new algorithm, \”scArches\”, which stands for single cell architecture surgery. The algorithm’s biggest advantage is that it uses transfer learning instead of raw data to compare single-cell genomics datasets with existing reference sets, preserving privacy and anonymity. The algorithm makes it easy to annotate and interpret new data sets and allows for a democratized use of single-cell atlases.
Source:
https://phys.org/news/2021-08-ai-diseased-cells.html
