ChIP-Seq/ATAC-Seq/CUT&RUN Analysis Report
2023-06-22
Figures are shown as examples only.
1. Introduction
ChIP-Seq is a method to study epigenetics
interaction between DNA and proteins by identifying the target binding
sites. For ChIP-Seq, we employ an in-house bioinformatics approach to
map reads, call peaks, perform differential analysis, and detect motifs
using reputable computational resources. In addition, we also offer
ATAC-Seq and CUT&RUN analysis. For
ATAC-Seq (Yan
et al., 2020) and CUT&RUN (Yu et
al., 2021) specifically, ENCODE ATAC-seq
pipeline and CUT&RUNTools 2.0
pipeline are proceeded to the peak detection step, respectively, and
then followed by our downstream analysis.
2. Bioinformatics Procedure and Data Processing
For our in-house bioinformatics pipeline, FastQC was used to check
the quality of raw and trimmed reads firstly. Trimmomatics was used to
cut adapters and trim low-quality bases with a default setting. After
mapping reads to the reference genome, mapped reads that have MAPQ score
< 10 were removed (filters may be set conditionally). Duplicates were
also removed. deepTools was used to normalized BAM and generate BW
format for visualization. MACS2 was used to call peak. Annotation was
performed using ChIPseeker (R package). If there was no replicate,
MAnorm (R package) was used for sample comparison. On the other hand, if
there were replicates, their called peaks were merged and DiffBind (R
package) was then used for the comparison. Other downstream procedures
included GO/KEGG (R package clusterProfiler), combined density profiles
(deepTools), and motif detection (MEME).
File
availability:
Raw / trimmed FastQ
Mapping statistics
FastQC
reports / MultiQC reports
BAM
BigWig (BW)
3. Peak Calling
Peak Calling is a computational method used to identify areas in the
genome that have been enriched with aligned reads. MACS algorithm can be
used for identifying transcription factor binding sites (narrow peaks)
and histone modification enriched regions (broad peaks). It outputs key
files such as peak files (file which contains the peak locations along
with peak summit, p-value, and q-value) and summit
files (file which contains peak summit locations for motif
analysis).
4. Clustering (DiffBind R package)
Sample Cluster Heatmap. Replicate samples.
Clustering of samples helps to segregate data into similar groups.
Correlation heatmap for all peak-calling samples.
PCA plot. Replicate samples. PCA is a procedure
which principle components are obtained by orthogonally transforming a
set of possibly correlated variables (high-dimensional data) into a
smaller number (few dimensions) of linearly uncorrelated variables.
Dimensional reduction of peak datasets using PCA.
5. Peak Overlay
Peak overlay plots show the number of peaks that are common and different between comparable datasets.
Venn diagram of binding site overlaps of replicate samples in the comparison groups.
6. Differential Analysis
Replicate samples. The differential analysis between replicate sample groups was performed using DiffBind (R package). This package includes functions that help to process peak sets, including merging overlapping peak sets in replicates, counting merged peak sets, and identifying statistically-significant differentially bound sites based on binding affinity between two groups. There are five outputs:I. MA plot
MA plot for data normalization. Red dots represent sites that are significantly differentially bound. p < 0.05. Blue line: 0-fold change.
II.Volcano plot
Volcano plot that shows significantly differentially bound sites. If log2(Target)-log2(Control) < 0, there are decreased affinity toward binding sties; whereas if log2(Target)-log2(Control) > 0, there are increased affinity toward binding sites.
III. Box plot
Box plot of read distributions for differentially bound sites.
IV. Heatmap
Heatmap between replicate samples of comparison groups.
(Table in CSV format)
Non-Replicate Samples. MAnorm method was used when there are non-replicate samples. ChIP-Seq data often have a lot of signal-to-background (S/N) ratios. This method focuses on the common sites and uses them as a reference for normalization.
MA plot before normalization using common peaks (left) and after normalization displaying all peaks (right). M: log2 fold changes. A: average expression signal. Green line: robust regression; red line: LOWESS regression; blue line: M = 0.
7. Functional Analysis
The function analysis helps to understand the functions and utilities of the biological systems. ChIPseeker (R package) was used for peak annotation. clusterProfiler (R package) was used for GO and KEGG-pathway enrichment analysis of individual sample peaks and comparison peaks.
Gene Ontology (GO) enrichment analysis. Dot plot of significant Biological Process GO terms.
KEGG pathway enrichment analysis. Bar graph of significant KEGG terms.
8. Combined Density Profile Analysis
Combined Density profile plots of peak regions were generated using deepTools.
Density plots and genomic heatmaps show respective peak intensities centered at the transcriptional start site (TSS) flanked by 3kb regions. x-axis represents the genomic location relative to the TSS. y-axis represents signal intensity.
9. Motif Analysis
MEME was used to analyze motif sequences around the peak regions. The
MEME-chIP performs motif discovery, enrichment, search, comparison, and
visualization.
Motif analysis. Top two motifs found within the peak regions (+100 nucleotides on each side) of a sample were shown. The first motif (E-value = 4.4e-182) was related to that of CTCF. The second motif (E-value = 1.9e-138) was related to that of SPIB or SPI1.
10. Citation
- Feng J, Liu T, and Zhang Y. 2011. Using MACS to identify peaks from ChIP-Seq data.Curr Protoc Bioinformatics. Chapter 2: Unit 2.14
- Kong NR, Chai L, Tenen DG, and Bassal MA. 2021. A modified CUT&RUN protocol and analysis pipeline to identify transcription factor binding sites in humna cell lines.STAR Protoc, 2(3):100750.
- Lever J, Krzywinski M, and Altman N. 2017. Principal component analysis. Nature Methods 14:641-642.
- Machanick P and Bailey TL. 2011. MEME-ChIP: motif analysis of large DNA datasets, Bioinformatics, 27(12):1696-1697.
- Ramirez F, Ryan DP, Gruning B, Bhardwaj V, Kilpert F, Richter AS, Heyne S, Dundar F, and Manke T. 2016. deepTools2: a next generation web server for deep-sequencing data analysis. Nucleic Acids Res, 44(W1):W160-165.
- R Development Core Team,
2008 R.
Allaire JJ et al., 2015. rmarkdown.
Sievert C et al., 2017 plotly.
Wickham H et al., 2009 ggplot2. - Shao Z, Zhang Y, Yuan GC, Orkin SH, and Waxman DJ. 2012. MAnorm: a robust model for quantitative comparison of ChIP-Seq data sets. Genome Biol, 13(3):R16.
- Yan
F, Powell DR, Curtis DJ, and Wong NC. 2020. From reads to insight: a
hitchhiker’s guide to ATAC-seq data analysis. Genome Biology,
21(1):22.
- Yu F, Sankaran VG, and Yuna GC. 2021. CUT&RUNTools 2.0: a pipeline for single-cell and bulk-level CUT&RUN and CUT&Tag data analysis. Bioinfomatics, 38(1):252-254.
- Yu G, Wang LG, and He QY. 2015. ChIPseeker: an R/Bioconductor package for ChIP peak annotation, comparison and visualization. Bioinformatics, 31(14):2382-2383.
Appendix
A list of software, tools, and libraries used in the analysis
pipeline:
Trimmomatic (v0.38)
FastQC (v0.11.8)
MultiQC
(v1.11)
SAMtools (v0.1.19)
Picard (v2.20.4)
bedTools
(v2.29.2)
deepTools (v3.4.1)
MACS (v2.2.6)
MEME (v5.1.1)
(https://meme-suite.org/meme/)
R libraries:
ChIPseeker (v1.22.1)
clusterProfiler (v3.14.3)
DiffBind (v3.8.4)
For ChIP-Seq:
BWA (v0.7.10)
For ATAC-Seq:
ENCODE ATAC-seq pipeline (https://github.com/ENCODE-DCC/atac-seq-pipeline)
Bowtie2 (v2.3.4.1)
For Cut&Run:
CUT&RUNTools 2.0 (https://github.com/fl-yu/CUT-RUNTools-2.0)
Bowtie2
(v2.3.4.1)
Language:
Shell
R (v4.3.0)
Python (v3.7.13)
Perl
(v5.26.2)
Abbreviations
ATAC-Seq Assay for Transposase-Accessible Chromatin with
Sequencing
ChIP-Seq Chromatin Immunoprecipitation with
Sequencing
CUT&RUN Cleavage Under Targets and Release Using
Nuclease
GO Gene Ontology
KEGG Kyoto Encyclopedia of Genes
and Genomes
MACS Model-based Analysis of ChIP-Seq
PCA
Principal Component Analysis