Cytometry Part AJClub
Citizen Science Gamers Enable Automated Flow Cytometry Gating Through Machine Learning.
Cytometry. Part A : the journal of the International Society for Analytical Cytology · · Journal Article · Open access
Montante, Yokosawa + more
Abstract ↗AI summary
The abstract is read at the publisher; the summary is JClub's.
FlowMagic trained on 31,703 citizen-science consensus annotations achieves 90% agreement for large and 65% for low-abundance immune populations on 92,203 flow cytometry plots.
- Why it matters: Automated flow cytometry gating is foundational yet subjective, and machine learning is limited by scarce high-quality labeled data needed for reproducible immunology analyses.
- What they did: Using EVE Online, 839,199 de-identified accounts generated 52,000 plots, which quality control distilled into 31,703 consensus annotations to train template and generalized gating models.
- The result: Template models match expert-provided gates best, while the generalized model enables gating without gate-specific templates but is constrained by missing expert-curated examples for complex, low-abundance, or context-dependent gates.
The findingWhy it mattersWhat they didThe result
AI summary of the abstract · may contain errors · not medical advice · read the paper.
- Open access