Differential Binding Analysis of ChIP-Seq peak data

Guide summary #

DiffBind: Differential binding analysis of ChIP-Seq peak data 2015-05-20

Introduction #

Identifying statistically significantly differentially bound sites based on evidence of binding affinity between two groups

Processing overview #

DiffBind involves five phases

1. Reading in peaksets: MACS와 같은 peak caller 프로그램의 결과 (eg. .csv, .xlsx)
2. Occupancy analysis: merge replicates or other peak caller result
3. Counting reads: using BAM files, make binding affinity matrix for each sample. With this matrix, samples can be re-clustered.
4. Differential binding affinity analysis: identify statiscally significiantly differentially bound sites between sample groups. using edgeR make p-value and FDR
5. Plotting and reporting: MA plot, correlation heatmap, PCA plots, Box plots

Example: obtaining differentially bound sites #

Dataset: ChIPs against the transcription factor ERa using 5 breast cancer cell lines (MCF-7 3 responsive to Tamoxifen, 2 resistant). There are at least 2 replicates. (Differential oestrogen receptor binding is associated with clinical outcome in breast cancer (Ross-Innes et al.) Nature, 2012-01-04)

Peak caller is MACS

Reading in the peaksets #

Counting reads #

Establishing a contrast #

Performaing the differential analysis #

Retrieving the differentially bound sites #

Example: plotting #

Venn diagrams #

MA plots #

PCA plots #

Boxplots #

Heatmaps #

Example: differential binding analysis using a blocking factor #

Example: occupancy analysis and overlaps #

Overlap rates #

Deriving consensus peaksets #

A complete occupancy analysis: identifying sites unique to a sample group #

Comparision of occupancy and affinity based analyses #

Technical notes #

Loading peaksets #

Merging peaks #

edgeR analysis #

DESeq analysis #

DESeq2 analysis #

Vignette Data #

Using DiffBind and ChIPQC together #

Suggested Pages #

• 0.025 Breast cancer
• 0.025 PR
• 0.025 January 25
• 0.025 Bowtie
• 0.025 FASTQ
• 0.025 UCSC
• 0.025 November 30
• 0.025 2018
• 0.025 Bioinformatics
• 0.025 Homebrew
• More suggestions...