Sequencing & Data Analysis

Overview

CUT&Tag sequencing data analysis transforms raw sequencing reads into biologically meaningful insights—identifying where specific proteins bind to the genome and how epigenetic modifications are distributed across chromatin. Compared to ChIP‑seq, CUT&Tag requires significantly fewer sequencing reads (3–8 million total reads, with 3–5 million unique reads typically sufficient) while delivering higher signal‑to‑noise ratios and lower background. The analysis workflow shares many steps with standard epigenomic pipelines but has specific considerations unique to the Tn5‑based tagmentation chemistry.

This guide outlines a complete data analysis workflow for CUT&Tag experiments, from quality control of raw FASTQ files through to peak calling, annotation, and visualization. It also provides quality control metrics—including FRiP scores, duplication rates, and TSS enrichment—to help you assess whether your CUT&Tag experiment has succeeded.

1. Quality Control and Preprocessing

1.1 Raw Data Quality Assessment

Raw sequencing data in FASTQ format should be assessed using FastQC to evaluate:

• Base Quality Distribution: Per‑base quality scores should remain high across all positions. Sequences with persistently low quality toward the 3' end should be trimmed or removed.
• Adapter Contamination: GC content distribution should be checked for abnormal peaks indicative of adapter contamination.
• Sequence Length Distribution: Short fragments introduced by sequencing truncation should be flagged.

MultiQC can aggregate FastQC reports across multiple samples to provide a consolidated visualization of quality metrics.

1.2 Adapter Trimming and Filtering

CUT&Tag libraries contain Illumina adapter sequences that must be removed before alignment. Cutadapt scans reads and removes the adapter wherever it occurs; it does not trim a fixed number of bases from the read start.

Adapter Sequence: The standard Illumina adapter sequence is AGATCGGAAGAGCACACGTCTGAACTCCAGTCAC.

Low‑Quality Filtering: Use Trimmomatic with a sliding window approach (4 bp window, average quality ≥15) to trim low‑quality bases.

Short Read Caution: While filtering reads below 36 bp is common, exercise caution—some transcription factor complexes produce genuine ultra‑short fragments (20–30 bp) that should not be discarded. Consider retaining all reads that pass quality thresholds and allowing the alignment step to filter unmappable reads.

Duplicate Removal: Use Picard MarkDuplicates to identify and remove PCR duplicates, retaining only uniquely aligned reads.

2. Sequence Alignment

2.1 Reference Genome Alignment

Aligned reads define the genomic coordinates of each sequenced fragment, enabling subsequent peak calling and annotation.

• Aligner: Bowtie2 is recommended in end‑to‑end mode (--very‑sensitive). Local alignment (--very‑sensitive‑local) is discouraged for CUT&Tag because Tn5 transposase cleavage produces precise fragment ends; soft‑clipping mismatches in local alignment can introduce false positives.
• Mismatch Settings: Use the Bowtie2 default ‑N 0 (no mismatches in the seed) to ensure high mapping specificity. Increasing to ‑N 1 offers no biological benefit and substantially increases alignment time and multi‑mapping errors.
• Genome Version: Use reference genomes consistent with the organism (e.g., hg38 for human, mm10 for mouse) and ensure the annotation file matches the genome build.

2.2 Alignment Quality Control

• Overall Alignment Rate: Should be ≥80%.
• Unique Mapping Rate: Reads mapping uniquely to the reference genome are essential for accurate peak calling.

If alignment rates are low, check genome version compatibility, adapter contamination, or sample cross‑contamination.

3. Peak Calling

Peak calling is the process of identifying genomic regions enriched for sequencing fragments relative to background, indicating protein‑DNA binding sites.

3.1 Peak Calling by Target Type

Target Type	Mode	Key Parameters	Notes
Histone Modifications (Broad)	Broad peak (--broad)	Refer to the recommended settings in your peak calling software manual	Captures broad, continuous modified domains (e.g., H3K27me3, H3K9me3)
Narrow Histone Marks	Narrow peak (--narrow)	Refer to the recommended settings in your peak calling software manual	For marks like H3K4me3 and H3K27ac that form focused peaks; parameters may vary by software and library preparation method
Transcription Factors	Narrow peak (--narrow)	Refer to the recommended settings in your peak calling software manual	TFs typically produce fragments <120 bp around binding sites; parameters may vary by software and sequencing strategy

3.2 Important parameter notes:

For CUT&Tag, transcription factors and narrow histone marks generally use a negative shift parameter to center reads on the Tn5 transposase cut sites. Positive shifts are typically artifacts of older single‑end ChIP‑seq workflows and are generally not recommended for CUT&Tag. Exact parameter values depend on the peak calling software and library preparation method; refer to the software manual for recommended settings.

3.3 Peak Calling Tools

A variety of peak calling tools are available for CUT&Tag data analysis. Commonly used options include MACS2, SEACR, and GoPeaks. The choice of tool and its parameters depends on the target type (e.g., broad vs. narrow peaks) and dataset characteristics. Users should consult the documentation of the selected tool for appropriate usage.

3.4 Control Samples

IgG control is the preferred negative control for CUT&Tag peak calling, as it accurately captures background signal from non‑specific antibody binding. Input DNA is less informative in CUT&Tag because Tn5 tagmentation kinetics dominate the background profile rather than random fragmentation.

4. Functional Annotation and Biological Interpretation

4.1 Peak Annotation

Annotation places peaks in their genomic context — promoters, gene bodies, enhancers, intergenic regions — linking binding events to potential target genes.

• Tools: Various R/Bioconductor and command‑line tools are available for peak annotation, such as ChIPseeker, HOMER, and ChIPpeakAnno. The choice of tool depends on the analysis workflow and data format.
• Promoter Range: The promoter region is typically defined as a fixed distance upstream and downstream of the TSS; the exact range may vary by mark type and study design. Narrow marks often use a tighter range than broad histone modifications.
• Annotation Outputs: Genomic region distribution (promoter, 5′ UTR, exon, intron, intergenic), nearest gene assignment, and distance to TSS.

4.2 Functional Enrichment Analysis

• Gene Ontology (GO): Biological process, molecular function, and cellular component enrichment.
• KEGG Pathways: Pathway enrichment analysis to identify regulated biological pathways.

For regulatory network analysis, tools such as GREAT can predict direct target genes by associating peaks with proximal and distal regulatory regions. Literature examples have demonstrated that certain histone modifications are enriched at promoters of key pathway genes, providing insight into pathway‑level regulatory mechanisms.

Key Quality Control Metrics

Metric	Description	Guidance
FRiP Score	Fraction of Reads in Peaks; measures the proportion of sequencing reads that fall within called peak regions	Generally higher for transcription factors than for histone marks; expected values vary by target and sample quality. A low FRiP may indicate poor antibody performance or insufficient sequencing depth.
TSS Enrichment	Enrichment of reads around TSS; reflects expected distribution of active marks	Active marks (e.g., H3K4me3, H3K27ac) typically show sharp, symmetrical enrichment around TSS; broad marks show wider distribution. Assess profile shape rather than relying on a single cutoff.
Replicate Concordance	Correlation between biological replicates	Higher correlation indicates better reproducibility; acceptable range depends on sample type and experimental variability.
Duplication Rate	Percentage of duplicate reads; CUT&Tag duplication rates vary by target type	Duplication rates are influenced by target abundance and sequencing depth rather than being a fixed indicator of data quality; assess in context of FRiP and enrichment.
Fragment Size Profile	Physical library fragment size distribution	Mononucleosome fragments (~270–300 bp including adapters) are typical for histone marks; broad marks may show dinucleosome peaks (~500 bp); transcription factors often show sub‑nucleosomal fragments (170–220 bp). Profile shape should be consistent with target type.
Promoter Region Percentage	Proportion of peaks located in promoter regions	Varies by target and cell type; active histone marks typically show substantial promoter localization, but exact proportions depend on biological context.

Notes:

• Duplication Rates: Duplication rates vary by target type and are influenced by genomic distribution. Narrow marks (e.g., H3K4me3) are restricted to promoter regions and saturate rapidly during sequencing, typically yielding higher duplication rates. Broad marks (e.g., H3K27me3) spread across large genomic domains, resulting in lower duplication rates. This reflects the expected biological pattern, not a technical artifact.
• Fragment Size: Core nucleosomal DNA is approximately 147 bp; standard Illumina adapters add additional sequence. Therefore, the physical mononucleosome library peak appears at a size consistent with the combined nucleosomal DNA and adapter lengths. Transcription factors typically show a dominant sub‑nucleosomal/nucleosome‑free fraction. The exact peak positions may vary depending on the specific library preparation kit and sequencing platform.
• TSS Enrichment: Active regulatory marks should display a sharp, symmetrical enrichment window flanking TSS. This is typically visualized as a meta‑gene plot or TSS enrichment heatmap.

6. Data Visualization

6.1 IGV Tracks

IGV (Integrative Genomics Viewer) is the standard tool for visualizing CUT&Tag signal tracks. What to expect:

• Positive Controls (H3K4me3): Sharp, focused peaks enriched at TSSs.
• Negative Controls (IgG): Flat, low‑signal baseline with no enrichment peaks.
• Treatment vs Control: Compare signal intensity between conditions; differences in peak height indicate changes in binding or modification levels.

6.2 Heatmaps and Meta‑Gene Plots

• Heatmaps: Display signal enrichment patterns across multiple genes; rows represent individual genes aligned at TSSs, columns represent genomic positions, and color intensity reflects signal strength. Red indicates high enrichment; blue indicates low enrichment.
• Meta‑Gene Plots: Average signal profiles across all genes aligned at TSS/TES; single curve showing overall enrichment pattern.

6.3 Genome Coverage Tracks

Generate normalized signal tracks (BigWig format) using deepTools, which can be visualized on genome browsers to provide a panoramic view of genome‑wide signal distribution.

7. Sequencing Depth and Read Metrics

7.1 Recommended Sequencing Depth

Parameter	Recommendation	Notes
Total Reads	Refer to the sequencing service provider or sequencing platform guidelines	Per library. Optimal depth depends on genome size, target abundance, and experimental objectives
Unique Reads	Sufficient for the specific analysis objectives	After filtering multi‑mapping and duplicate reads. Required number varies with genome size and target type
Deeper Sequencing	May be considered for rare targets or low‑input samples	Increases coverage but may also increase duplicate rates; cost‑benefit should be evaluated case‑by‑case.

Summary

CUT&Tag data analysis follows a structured workflow from raw reads to biological interpretation:

1. Quality Control: Assess raw read quality, remove adapters (by scanning, not fixed trimming), and filter low‑quality bases cautiously to retain genuine short fragments.
2. Alignment: Map reads to the reference genome using a suitable aligner (e.g., Bowtie2) with appropriate parameters for CUT&Tag data.
3. Peak Calling: Use peak calling software appropriate for the target type (broad or narrow mode). For CUT&Tag data, negative shift parameters may be applied depending on the software and target; refer to the software manual for recommended settings.
4. Annotation: Assign peaks to genomic features and nearest genes using standard annotation tools.
5. Quality Assessment: Evaluate FRiP scores, duplication rates, TSS enrichment, and replicate concordance against expected patterns for the target type.
6. Visualization: Generate IGV tracks, heatmaps, and meta‑gene plots using visualization tools such as deepTools and IGV.

Proper quality control at each step ensures that downstream biological interpretations are reliable. Regular updates to genome annotation files are recommended for consistent data quality.