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

Area of Science:

• Genomics
• Epigenetics
• Bioinformatics

Background:

• Assay for transposase-accessible chromatin via sequencing (ATAC-seq) is crucial for studying genome accessibility.
• Detecting transcription factor (TF) binding and nucleosome occupancy using ATAC-seq is hindered by sequence-dependent Tn5 cleavage bias.
• This bias varies significantly between samples, complicating accurate data interpretation.

Purpose of the Study:

• To develop a robust framework for accurate ATAC-seq footprinting by addressing Tn5 cleavage bias.
• To enhance the reliability of TF and nucleosome occupancy detection in both bulk and single-cell ATAC-seq data.
• To improve downstream single-cell analyses, including TF activity inference and cell age prediction.

Main Methods:

• Introduced TraceBind, an ATAC-seq footprinting framework.
• Implemented sample-specific Tn5 bias correction by fine-tuning a pretrained cleavage model using mitochondrial DNA reads.
• Employed multiscale detection with empirical false discovery rate control.

Main Results:

• TraceBind effectively reduces false positives in ATAC-seq data while maintaining high sensitivity.
• Validated performance across bulk and single-cell ATAC-seq datasets, naked DNA controls, ChIP-seq, CUT&RUN, and TF degradation perturbations.
• Demonstrated improved downstream single-cell analyses, including TF activity inference and cell age prediction.

Conclusions:

• TraceBind provides a significant advancement in ATAC-seq data analysis by correcting for Tn5 cleavage bias.
• The framework enhances the accuracy of TF binding and nucleosome occupancy detection.
• TraceBind reveals regulatory information beyond DNA sequence alone, complementing DNA foundation models.