Semi-supervised peak calling with SPAN and JBR genome browser.

Area of Science:

• Genomics
• Bioinformatics
• Computational Biology

Background:

• ChIP-seq and ATAC-seq experiments generate multiple epigenetic profiles, often with replicates.
• Analyzing multiple samples presents challenges in peak calling consistency and preserving individual sample statistics.
• Existing methods struggle with independent peak calling leading to variable peak numbers or joint analysis averaging out individual differences.

Purpose of the Study:

• To present the implementation of the SPAN (semi-supervised peak calling) approach for robust multi-sample epigenetic data analysis.
• To introduce the JBR genome browser as a stand-alone tool for streamlined annotation, analysis, and visualization of epigenetic profiles.
• To provide a solution for consistent group-level analysis while maintaining individual sample statistics.

Main Methods:

• Implementation of a semi-supervised peak calling approach (SPAN).
• Development of the JBR genome browser for interactive manual region selection and annotation.
• Integration of peak optimization for calibrating sample-specific peak calling parameters using manual annotations.
• Design for efficient data processing to handle thousands of tracks.

Main Results:

• SPAN enables robust analysis of multiple epigenetic profiles, preserving individual sample statistics.
• JBR facilitates accessible and streamlined annotation, analysis, and visualization of ChIP-seq and ATAC-seq data.
• The combined approach addresses challenges in multi-sample epigenetic data analysis, improving consistency.
• JBR supports efficient processing, enabling fast viewing and analysis of up to thousands of tracks.

Conclusions:

• JBR and SPAN offer next-generation tools for multi-sample epigenetic data visualization and analysis.
• The implementation provides a solution for consistent and robust analysis of diverse ChIP-seq and ATAC-seq datasets, including single-cell experiments.
• This approach enhances the ability to perform group-level analysis without losing individual sample information.