CHIP-SEQ后期数据解析

第一部分：chip-seq+rna-seq

这里详细研究一下Rcade: R-based analysis of ChIP-seq And DifferentialExpression data

Rcade is a bioconductor tool that analyses ChIP-seq data by coupling the ChIP-seq resultsto an existing Differential Expression(DE) analysis. Rcade enables inferring direct targets of a transcription factor (TF) - thesetargets should exhibit TF binding activity, and their expression levels shouldchange in response to TF binding to gene associated regulatory elements.

1.      安装相关包

Rgl，biomaRt，Rcade

2.      Data地址

library("Rcade")

dir <- file.path(system.file("extdata", package="Rcade"), "STAT1")

dir

3.      DE analysis. 建议使用limm处理差异基因。

DE <- read.csv(file.path(dir,"DE.csv"))

DElookup <-list(GeneID="ENSG", logFC="logFC",B="B","Genes.Location", "Symbol")

4.       Chip-seq analysis

dir(dir, pattern = ".bam")

targets <- read.csv(file.path(dir,"targets.csv"), as.is = TRUE)

targets

5.       Annotation

anno <- read.csv(file.path(dir,"anno.csv"))

anno <- anno[order(anno$chromosome_name),]

colnames(anno) <-c("ENSG","chr","start","end","str")

ChIPannoZones <-defineBins(anno, zone=c(-1500, 1500), geneID="ENSG")

6.       Prior specification

DE.prior = 0.01

prior.mode = "keepChIP"

prior = c("D|C" = 0.05, "D|notC" = 0.005)

7.      Analysis

UseRcadeAnalysis function:

Rcade <-RcadeAnalysis(DE, ChIPannoZones,annoZoneGeneidName="ENSG", ChIPtargets=targets,ChIPfileDir = dir, DE.prior=DE.prior,

prior.mode=prior.mode,prior=prior,DElookup=DElookup)

Rcade

The Rcade object stores information from the analysis. The DE, ChIP andRcade analysis can be then stored in seperate objects.

xDE <-getDE(Rcade)

xChIP <-getChIP(Rcade)

xRcade <- getRcade(Rcade)

8.       Plottingand QC

1. PrincipleComponent Analysis on the counts with the plotPCA function

plotPCA(Rcade)

2. The MMplot shows log-ratios from DE plotted against log-ratios from the ChIP-seq

plotPCA(Rcade)

plotMM(Rcade)

3. PlotBBB() plots log-odds values forChIP-seq, DE and the combined ChIP-seq/DE analysis. Youneed to install the rglpackage.

library(rgl)

plotBBB(Rcade)

Exportresults：

exportRcade(Rcade,directory="RcadeOutput",cutoffArg=2000)

moreadvanced export options are available:

exportRcade(Rcade,directory="RcadeOutput", cutoffMode="top", cutoffArg =1000,

justGeneID=FALSE,removeDuplicates="beforeCutoff")

Exportfiles

This function exportsRcade output to disk - speci_cally, it creates the following _les:

_ ChIP.csv

_ ChIPonly.csv

_ DEandChIP.csv

_ DownChIP.csv

_ Down.csv

_ DownNoChIP.csv

_ Nothing.csv

_ UpChIP.csv

_ Up.csv

_ UpNoChIP.csv

Each _le containsgenes appropriate to its hypothesis, sorted by descending B value (i.e. rankedfrom

most interesting toleast interesting). For example, if you wanted the genes that display DE(either up

or down) and also haveChIP signal present, you would look at the top rows of DEandChIP.csv. For

genes that have a ChIPsignal but explicitly show no DE, use ChIPonly.csv.

Rcade is an extremelyuseful package for downstream analysis of ChIP-seq data. It goes beyond peak

or gene lists andenables identi_cation of direct targets of transcription factors. These can beused in

combination withNetwork Biology methods to unravel complexities of biological systems.