Effects of Distance Metrics and Scaling on the Perturbation Discrimination Score

2025/11/21 by Qiyuan Liu, Qirui Zhang, Liu, Qiyuan +7
Biochemistry, Genetics and Molecular Biology · #Applications (stat.AP) #Bioinformatics and Genomic Networks #Computation (stat.CO) #FOS: Computer and information sciences #Gene Regulatory Network Analysis #Single-cell and spatial transcriptomics

paper · pdf · doi:10.48550/arxiv.2511.16954

openalex publication_date 2025/11/21 · openalex created_date 2025/11/25 · openalex updated_date 2026/07/28

Abstract

The Perturbation Discrimination Score (PDS) is increasingly used to evaluate whether predicted perturbation effects remain distinguishable, including in Systema and the Virtual Cell Challenge. However, its behavior in high-dimensional gene-expression settings has not been examined in detail. We show that PDS is highly sensitive to the choice of similarity or distance measure and to the scale of predicted effects. Analysis of observed perturbation responses reveals that ℓ₁ and ℓ₂-based PDS behave very differently from cosine-based measures, even after norm matching. We provide geometric insight and discuss implications for future discrimination-based evaluation metrics.

Citations

• PerturBench: Benchmarking Machine Learning Models for Cellular Perturbation Analysis

Related

• Diversity by Design: Addressing Mode Collapse Improves scRNA-seq Perturbation Modeling on Well-Calibrated Metrics
• PerturBench: Benchmarking Machine Learning Models for Cellular Perturbation Analysis
• Improving Accuracy in Cell-Perturbation Experiments by Leveraging Auxiliary Information
• On Selecting Distance Metrics in n-Dimensional Normed Vector Spaces of Cells: A Novel Criterion and Similarity Measure Towards Efficient and Accurate Omics Analysis
• Effect Size Estimation and Misclassification Rate Based Variable Selection in Linear Discriminant Analysis