HybridQC: Machine Learning-Augmented Quality Control for Single-Cell RNA-seq Data

TL;DR

HybridQC is an R package that enhances single-cell RNA-seq data quality control by integrating traditional filtering with machine learning outlier detection, improving accuracy and adaptability for diverse datasets.

Contribution

It introduces a novel, lightweight framework combining threshold-based filtering with machine learning methods like Isolation Forest for improved QC in scRNA-seq data.

Findings

Effective identification of low-quality cells in noisy datasets

Compatibility with common formats like Seurat objects

Suitable for small-to-medium research projects

Abstract

HybridQC is an R package that streamlines quality control (QC) of single-cell RNA sequencing (scRNA-seq) data by combining traditional threshold-based filtering with machine learning-based outlier detection. It provides an efficient and adaptive framework to identify low-quality cells in noisy or shallow-depth datasets using techniques such as Isolation Forest, while remaining compatible with widely adopted formats such as Seurat objects. The package is lightweight, easy to install, and suitable for small-to-medium scRNA-seq datasets in research settings. HybridQC is especially useful for projects involving non-model organisms, rare samples, or pilot studies, where automated and flexible QC is critical for reproducibility and downstream analysis.