Reconstruction Set Test (RESET): a computationally efficient method for single sample gene set testing based on...

Area of Science:

• Bioinformatics
• Computational Biology
• Genomics

Background:

• Current single-sample gene set testing methods primarily detect mean differences and struggle with differential correlation patterns.
• Existing techniques face limitations in detecting competitive gene set scenarios and comparing scores across different gene sets or samples.
• Computational demands of existing methods can be prohibitive for large-scale genomic datasets.

Approach:

• Developed Reconstruction Set Test (RESET), a novel single-sample gene set testing method.
• Utilizes a computationally efficient randomized reduced rank reconstruction algorithm.
• Quantifies gene set importance by assessing the ability of set genes to reconstruct values for all measured genes.

Key Points:

• RESET effectively detects differential abundance and correlation patterns in both self-contained and competitive scenarios.
• Offers accurate score comparison across samples for a single gene set and between different gene sets.
• Demonstrates superior accuracy and lower computational cost compared to existing single-sample methods using scRNA-seq data.

Conclusions:

• RESET provides a powerful and versatile tool for gene set analysis in single samples.
• Addresses key limitations of current methods, enabling more robust and efficient genomic data interpretation.
• Facilitates advanced analysis of gene expression patterns, particularly for large and complex datasets.