Implicit reconstruction of gating strategies from sample-level flow cytometry annotations
- 1Helmholtz Munich
Abstract
Flow cytometry is central to clinical hematology, generating extensive archives comprising billions of single-cell profiles linked to expert-derived diagnostic summaries. While these records represent a largely untapped repository of collective expert knowledge, the underlying single-cell annotations required for supervised machine learning are typically not recorded in a standardized, accessible format. We present IMPGAT (implicit gating), a weakly supervised multiple-instance learning framework that reconstructs single-cell gating strategies directly from existing sample-level cell-type frequency distributions. The model utilizes an induced set attention mechanism to map single-cell marker profiles to specific cell populations, requiring neither explicit cell-level labels nor unsupervised clustering during training. We evaluate the framework on acute lymphoblastic leukemia (ALL) and acute myeloid leukemia (AML) cohorts, totaling 711 measurements from 268 individuals and 240 million single cells. In ALL, the models quantification of cell-type abundances shows no statistically significant difference from human interexpert variability. In AML, the framework achieves blast quantification on par with a recently established supervised state-of-the-art model that relies on explicit single-cell annotations for training. Error analysis shows that mismatches between model predictions and expert gating primarily involve expert-labeled non-cellular events, reflecting the non-incorporation of any explicit physical prior knowledge (such as the geometric relationships traditionally used to identify doublets and debris in scatter cytograms) in model training. Conversely, model certainty correlates with phenotypic borderline cases where manual bisection is inherently difficult. Our approach enables the systematic utilization of retrospective clinical records for the development of automated, standardized gating logic and is applicable to broader high-throughput contexts, such as large-scale immune monitoring or phenotypic drug screening.
Related
- Free flap selection strategies and advances in facial soft tissue defect reconstruction
- Operative management of abdominal bulge following deep inferior epigastric perforator flap breast reconstruction: a systematic review of repair strategies and outcomes
- Quadriceps Strength After Anterior Cruciate Ligament Reconstruction: Mechanisms of Weakness and Evidence-Based Rehabilitation Strategies
- Surface engineering strategies for PET-Based ligament scaffolds in anterior cruciate ligament reconstruction: a systematic review of mechanical and biological performances.
- Tissue engineering strategies from regional functional units to integrated functional reconstruction for intervertebral disc regeneration
BibTeX
@misc{buser2026implicit,
title = {{Implicit reconstruction of gating strategies from sample-level flow cytometry annotations}},
author = {M. Buser and M. Szczepanowski and A. Laqua and G. Chitadze and C. Baldus and M. Brueggemann and C. Marr},
howpublished = {medRxiv},
year = {2026},
month = oct,
doi = {10.64898/2026.10.20.26365931},
url = {https://www.medrxiv.org/content/10.64898/2026.10.20.26365931v1},
note = {CC BY-NC},
}
RIS
TY - UNPB
TI - Implicit reconstruction of gating strategies from sample-level flow cytometry annotations
AU - M. Buser
AU - M. Szczepanowski
AU - A. Laqua
AU - G. Chitadze
AU - C. Baldus
AU - M. Brueggemann
AU - C. Marr
PY - 2026
DA - 2026/10/20
PB - medRxiv
DO - 10.64898/2026.10.20.26365931
UR - https://www.medrxiv.org/content/10.64898/2026.10.20.26365931v1
AB - Flow cytometry is central to clinical hematology, generating extensive archives comprising billions of single-cell profiles linked to expert-derived diagnostic summaries. While these records represent a largely untapped repository of collective expert knowledge, the underlying single-cell annotations required for supervised machine learning are typically not recorded in a standardized, accessible format. We present IMPGAT (implicit gating), a weakly supervised multiple-instance learning framework that reconstructs single-cell gating strategies directly from existing sample-level cell-type frequency distributions. The model utilizes an induced set attention mechanism to map single-cell marker profiles to specific cell populations, requiring neither explicit cell-level labels nor unsupervised clustering during training. We evaluate the framework on acute lymphoblastic leukemia (ALL) and acute myeloid leukemia (AML) cohorts, totaling 711 measurements from 268 individuals and 240 million single cells. In ALL, the models quantification of cell-type abundances shows no statistically significant difference from human interexpert variability. In AML, the framework achieves blast quantification on par with a recently established supervised state-of-the-art model that relies on explicit single-cell annotations for training. Error analysis shows that mismatches between model predictions and expert gating primarily involve expert-labeled non-cellular events, reflecting the non-incorporation of any explicit physical prior knowledge (such as the geometric relationships traditionally used to identify doublets and debris in scatter cytograms) in model training. Conversely, model certainty correlates with phenotypic borderline cases where manual bisection is inherently difficult. Our approach enables the systematic utilization of retrospective clinical records for the development of automated, standardized gating logic and is applicable to broader high-throughput contexts, such as large-scale immune monitoring or phenotypic drug screening.
ER -