Robust footprinting with sample-specific Tn5 bias correction for bulk and single cell ATAC-seq

Area of Science:

• Genomics
• Computational Biology
• Epigenetics

Background:

• Assay for transposase-accessible chromatin via sequencing (ATAC-seq) provides base-resolution mapping of regulatory elements.
• Data sparsity and Tn5 transposase cleavage bias limit ATAC-seq power and introduce false positives.
• Sample-to-sample variability in Tn5 bias represents an underappreciated source of batch effects.

Purpose of the Study:

• To develop a computational framework, TraceBIND, for correcting sample-specific Tn5 bias in ATAC-seq data.
• To improve the identification of transcription factor (TF) and nucleosome footprints.
• To enhance the discovery of regulatory signals in single-cell ATAC-seq (scATAC-seq) data.

Main Methods:

• Developed TraceBIND, an unsupervised computational framework for Tn5 bias correction.
• Employed a dynamic flanking window statistical scan to identify TF and nucleosome footprints.
• Validated TraceBIND using multiple analyses, including comparison with supervised methods and application to scATAC-seq data.

Main Results:

• TraceBIND significantly reduces false discoveries and controls type 1 error while maintaining high sensitivity.
• The framework demonstrates comparable power to supervised methods for detecting known TF binding sites without requiring ChIP-seq data.
• TraceBIND identified age-associated regulatory changes in rat kidney scATAC-seq data, linking footprint activity to epigenetic drift.

Conclusions:

• TraceBIND effectively corrects sample-specific Tn5 bias, improving ATAC-seq data reliability.
• Footprint-informed scATAC-seq analysis using TraceBIND reveals regulatory signals missed by conventional peak-based methods.
• This approach offers a powerful tool for understanding epigenetic regulation and age-associated changes.