A biology-driven deep generative model for cell-type annotation in cytometry

TL;DR

This paper introduces Scyan, a deep generative model that automates cell-type annotation in cytometry data, improving accuracy, speed, and interpretability over manual gating and existing models, while also handling batch effects and other tasks.

Contribution

The paper presents Scyan, a novel biology-driven deep generative model that automates cell-type annotation in high-dimensional cytometry data, outperforming existing methods and addressing batch effects.

Findings

Scyan outperforms state-of-the-art models on multiple datasets.

Scyan is faster and more interpretable than existing methods.

Scyan effectively handles batch-effect removal and other tasks.

Abstract

Cytometry enables precise single-cell phenotyping within heterogeneous populations. These cell types are traditionally annotated via manual gating, but this method suffers from a lack of reproducibility and sensitivity to batch-effect. Also, the most recent cytometers - spectral flow or mass cytometers - create rich and high-dimensional data whose analysis via manual gating becomes challenging and time-consuming. To tackle these limitations, we introduce Scyan (https://github.com/MICS-Lab/scyan), a Single-cell Cytometry Annotation Network that automatically annotates cell types using only prior expert knowledge about the cytometry panel. We demonstrate that Scyan significantly outperforms the related state-of-the-art models on multiple public datasets while being faster and interpretable. In addition, Scyan overcomes several complementary tasks such as batch-effect removal, debarcoding, and population discovery. Overall, this model accelerates and eases cell population characterisation, quantification, and discovery in cytometry.