SPARCS: AI analyzes 70 million cells at once to mine gene function at scale
(Team at Ludwig Maximilian University of Munich, paper published in Cell, 2026-10-08)
Key points
- What SPARCS is: a new technique that can recover complex image phenotypes at scale — it uses AI to analyze microscope images and systematically screen how gene alterations affect cells, turning "finding gene function by looking at images" into a genome-wide undertaking.
- Scale: microscope images of 70 million cells analyzed in a single run.
- Proof of concept: in a genome-wide experiment on autophagy (an important cellular recycling process), SPARCS identified most of the known genes that regulate autophagosome formation and also found new participating genes.
- A new immune mechanism: in studying STING (a key sensor of the innate immune system), the team found that the acidity of the Golgi apparatus is critical for its early trafficking, and that the GPHR protein regulates this immune sensor by affecting pH.
- A closed loop: the selected cells are fully isolated for mass spectrometry analysis, precisely characterizing the effects of the gene alterations at the molecular level — AI imaging screens and proteomics corroborate each other.
Source: read the original. This text is a summary for reference only.
