single-cell sequencing

A new graph-attention based AI method can generate synthetic representations of rare cell types, addressing a longstanding gap in single-cell RNA sequencing analysis where scarce cell populations are hard to study. By augmenting datasets with these generated cells, researchers aim to improve the accuracy of cell-type identification and downstream biological interpretation. Coverage so far is limited to a single technology news outlet, and details about the team, method validation, or publication venue are not yet available.

A report in Nature describes progress in mapping sugars at the level of individual cells. Unlike DNA or proteins, glycans on cell surfaces are hard to sequence, and new methods aim to detect which sugar structures each cell carries. Researchers say this could clarify how cells communicate and how diseases like cancer alter their surface chemistry.