# Robust and efficient annotation of cell states through gene signature scoring

## TL;DR

This paper introduces a new method for analyzing single-cell RNA sequencing data that improves the accuracy of identifying cell states.

## Contribution

The novel Adjusted Neighborhood Scoring (ANS) algorithm enhances score stability and cross-signature comparability for cell-state annotation.

## Key findings

- Existing gene signature scoring methods like Seurat, SCANPY, UCell, and JASMINE show insufficient performance for robust cell-state annotation.
- Adjusted Neighborhood Scoring (ANS) achieves comparable accuracy to supervised methods while being deterministic and more stable.
- ANS was successfully applied to distinguish cancer-associated fibroblasts from epithelial-to-mesenchymal transition cells.

## Abstract

Gene signature scoring is integral to single-cell RNA sequencing (scRNA-seq) data analysis, particularly for unsupervised cellular state annotation based on maximum signature score values. However, this application requires robust and comparable score distributions across diverse signatures and experimental conditions. Our systematic evaluation of established scoring methodologies—Seurat, SCANPY, UCell, and JASMINE—across nine healthy and cancer scRNA-seq data sets demonstrates their insufficiency in fulfilling this requirement. To address this limitation, we present Adjusted Neighborhood Scoring (ANS), a deterministic algorithm with enhanced control gene selection that significantly improves score stability and cross-signature comparability, achieving cell-state annotation accuracy comparable to supervised methods. We demonstrate the practical utility of ANS by developing and validating a gene signature to differentiate cancer-associated fibroblasts from malignant cells undergoing epithelial-to-mesenchymal transition. Overall, ANS provides a robust and reliable gene signature scoring framework, significantly improving the accuracy of score-based annotation of cell types and states in single-cell studies.