Comprehensive noise reduction in single-cell data with the RECODE platform.

Area of Science:

• Genomics
• Computational Biology
• Bioinformatics

Background:

• Single-cell sequencing offers deep biological insights but suffers from technical noise and batch effects.
• These artifacts obscure cellular structures, complicating rare cell type identification and multi-dataset comparisons.
• Existing methods struggle to address noise comprehensively across diverse single-cell data types.

Purpose of the Study:

• To upgrade the RECODE tool for simultaneous reduction of technical and batch noise in single-cell data.
• To extend the applicability of RECODE to various single-cell modalities beyond RNA sequencing.
• To enhance the accuracy and computational efficiency of noise mitigation techniques.

Main Methods:

• Development of iRECODE, an extension of the RECODE algorithm, for joint noise reduction.
• Adaptation of the RECODE platform for single-cell high-throughput chromosome conformation capture (Hi-C) and spatial transcriptomics data.
• Algorithmic improvements to enhance noise reduction accuracy and computational performance.

Main Results:

• iRECODE effectively reduces both technical and batch noise in single-cell datasets.
• The enhanced RECODE platform demonstrates robust performance across transcriptomic, epigenomic, and spatial data.
• Significant improvements in computational efficiency and analytical accuracy were achieved.

Conclusions:

• The upgraded RECODE platform, including iRECODE, provides a versatile and powerful solution for noise mitigation in single-cell analyses.
• This tool enables more reliable detection of rare cell types and facilitates accurate cross-dataset comparisons.
• The enhanced RECODE facilitates deeper biological insights from diverse single-cell data modalities.