Automated annotation of rare-cell types from single-cell RNA-sequencing data through synthetic oversampling

Abstract

The research landscape of single-cell and single-nuclei RNA sequencing is evolving rapidly, and one area that is enabled by this technology, is the detection of rare cells. An automated, unbiased and accurate annotation of rare subpopulations is challenging. Once rare cells are identified in one dataset, it will usually be necessary to generate other datasets to enrich the analysis (e.g., with samples from other tissues). From a machine learning perspective, the challenge arises from the fact that rare cell subpopulations constitute an imbalanced classification problem. We here introduce a Machine Learning (ML)-based oversampling method that uses gene expression counts of already identified rare cells as an input to generate synthetic cells to then identify similar (rare) cells in other publicly available experiments. We utilize single-cell synthetic oversampling (sc-SynO), which is based on the Localized Random Affine Shadowsampling (LoRAS) algorithm. The algorithm corrects for the overall imbalance ratio of the minority and majority class. We demonstrate the effectiveness of the method for two independent use cases, each consisting of two published datasets. The first use case identifies cardiac glial cells in snRNA-Seq data (17 nuclei out of 8,635). This use case was designed to take a larger imbalance ratio ([~]1 to 500) into account and only uses single-nuclei data. The second use case was designed to jointly use snRNA-Seq data and scRNA-Seq on a lower imbalance ratio ([~]1 to 26) for the training step to likewise investigate the potential of the algorithm to consider both single cell capture procedures and the impact of "less" rare-cell types. For validation purposes, all datasets have also been analyzed in a traditional manner using common data analysis approaches, such as the Seurat3 workflow. Our algorithm identifies rare-cell populations with a high accuracy and low false positive detection rate. A striking benefit of our algorithm is that it can be readily implemented in other and existing workflows. The code basis is publicly available at FairdomHub (https://fairdomhub.org/assays/1368) and can easily be transferred to train other customized approaches.