ChIP-seq详细分析流程

参考生信技能树1以及生信技能树2
只记录从数据下载，到最终结果展示，具体生物学知识请自行查阅
稍后关于ChIP-seq的背景知识我会再发布一篇文章。
数据下载：数据存放地址
关于环境配置和软件安装请参考我之前的RNA-seq分析流程

-------------------------------------------

1 软件安装及环境配置

2 sra数据下载(采取aspera下载）

cd /mnt/f/Data/chip_seq/data
for ((i=204;i<=209;i++));do ascp -QT -v -i ~/.aspera/connect/etc/asperaweb_id_dsa.openssh -k 1 -T -l200m [email protected]:/sra/sra-instant/reads/ByRun/sra/SRR/SRR620/SRR620${i}/SRR620${i}.sra .;done

ls -lh *sra
-rwxrwxrwx 1 root root 522M Aug 14 10:03 SRR620204.sra
-rwxrwxrwx 1 root root 365M Aug 14 10:08 SRR620205.sra
-rwxrwxrwx 1 root root 572M Aug 14 10:19 SRR620206.sra
-rwxrwxrwx 1 root root 741M Aug 14 10:30 SRR620207.sra
-rwxrwxrwx 1 root root 346M Aug 14 10:37 SRR620208.sra
-rwxrwxrwx 1 root root 639M Aug 14 11:08 SRR620209.sra

3 sra到fastq格式转换并进行质量控制

3.1 用samtools中的fastq-dump将sra格式转为fastq格式

source ~/miniconda3/bin/activate
for ((i=204;i<=209;i++));do fastq-dump --gzip --split-3 -A SRR620$i -O .;done

ls -lh *fastq.gz
-rwxrwxrwx 1 root root 883M Aug 14 11:18 SRR620204.fastq.gz
-rwxrwxrwx 1 root root 703M Aug 14 11:24 SRR620205.fastq.gz
-rwxrwxrwx 1 root root 968M Aug 14 11:30 SRR620206.fastq.gz
-rwxrwxrwx 1 root root 1.4G Aug 14 11:41 SRR620207.fastq.gz
-rwxrwxrwx 1 root root 589M Aug 14 11:45 SRR620208.fastq.gz
-rwxrwxrwx 1 root root 1.1G Aug 14 11:54 SRR620209.fastq.gz

3.2 用fastqc进行质量控制

#将所有的数据进行质控，得到zip的压缩文件和html文件
fastqc -o .  *.fastq.gz

4 Bowtie2进行序列比对

下载小鼠index文件并解压（迅雷下载）

unzip mm10.zip

unzip mm10.zip
Archive:  mm10.zip
  inflating: mm10.1.bt2
  inflating: mm10.2.bt2
  inflating: mm10.3.bt2
  inflating: mm10.4.bt2
  inflating: mm10.rev.1.bt2
  inflating: mm10.rev.2.bt2
  inflating: make_mm10.sh

cd /mnt/f/Data/chip_seq/data
bowtie2 -p 6 -3 5 --local -x /mnt/f/Data/reference/bowtie2_index/mm10 -U SRR620204.fastq.gz | samtools sort -O bam -o ../aligned/ring1B.bam &

19687027 reads; of these:
  19687027 (100.00%) were unpaired; of these:
    7778437 (39.51%) aligned 0 times
    7766495 (39.45%) aligned exactly 1 time
    4142095 (21.04%) aligned >1 times
60.49% overall alignment rate
[bam_sort_core] merging from 4 files and 1 in-memory blocks...

bowtie2 -p 6 -3 5 --local -x /mnt/f/Data/reference/bowtie2_index/mm10 -U SRR620205.fastq.gz | samtools sort -O bam -o ../aligned/cbx7.bam &

20710168 reads; of these:
  20710168 (100.00%) were unpaired; of these:
    2584870 (12.48%) aligned 0 times
    10914615 (52.70%) aligned exactly 1 time
    7210683 (34.82%) aligned >1 times
87.52% overall alignment rate
[bam_sort_core] merging from 4 files and 1 in-memory blocks...

bowtie2 -p 6 -3 5 --local -x /mnt/f/Data/reference/bowtie2_index/mm10 -U SRR620206.fastq.gz | samtools sort -O bam -o ../aligned/suz12.bam &

21551927 reads; of these:
  21551927 (100.00%) were unpaired; of these:
    7109852 (32.99%) aligned 0 times
    9455003 (43.87%) aligned exactly 1 time
    4987072 (23.14%) aligned >1 times
67.01% overall alignment rate
[bam_sort_core] merging from 4 files and 1 in-memory blocks...

bowtie2 -p 6 -3 5 --local -x /mnt/f/Data/reference/bowtie2_index/mm10 -U SRR620207.fastq.gz | samtools sort -O bam -o ../aligned/RYBP.bam &

39870646 reads; of these:
  39870646 (100.00%) were unpaired; of these:
    6312903 (15.83%) aligned 0 times
    20702502 (51.92%) aligned exactly 1 time
    12855241 (32.24%) aligned >1 times
84.17% overall alignment rate
[bam_sort_core] merging from 8 files and 1 in-memory blocks...

bowtie2 -p 6 -3 5 --local -x /mnt/f/Data/reference/bowtie2_index/mm10 -U SRR620208.fastq.gz | samtools sort -O bam -o ../aligned/IgGold.bam &

13291429 reads; of these:
  13291429 (100.00%) were unpaired; of these:
    5608796 (42.20%) aligned 0 times
    4663584 (35.09%) aligned exactly 1 time
    3019049 (22.71%) aligned >1 times
57.80% overall alignment rate
[bam_sort_core] merging from 2 files and 1 in-memory blocks...

bowtie2 -p 6 -3 5 --local -x /mnt/f/Data/reference/bowtie2_index/mm10 -U SRR620209.fastq.gz | samtools sort -O bam -o ../aligned/IgG.bam &

31218866 reads; of these:
  31218866 (100.00%) were unpaired; of these:
    5370101 (17.20%) aligned 0 times
    15180537 (48.63%) aligned exactly 1 time
    10668228 (34.17%) aligned >1 times
82.80% overall alignment rate
[bam_sort_core] merging from 6 files and 1 in-memory blocks...

cd ../aligned
ls -lh *bam

-rwxrwxrwx 1 root root 623M Aug 14 23:53 cbx7.bam
-rwxrwxrwx 1 root root 999M Aug 14 23:56 IgG.bam
-rwxrwxrwx 1 root root 503M Aug 14 23:43 IgGold.bam
-rwxrwxrwx 1 root root 752M Aug 14 19:37 ring1B.bam
-rwxrwxrwx 1 root root 1.2G Aug 14 23:59 RYBP.bam
-rwxrwxrwx 1 root root 845M Aug 14 23:49 suz12.bam

5 用MACS2获取Chip-seq富集区

macs2 callpeak -c IgGold.bam -t suz12.bam -q 0.05 -f BAM -g mm -n suz12 &
macs2 callpeak -c IgGold.bam -t cbx7.bam -q 0.05 -f BAM -g mm -n cbx7 &
macs2 callpeak -c IgGold.bam -t ring1B.bam -q 0.05 -f BAM -g mm -n ring1B &
macs2 callpeak -c IgGold.bam -t RYBP.bam -q 0.05 -f BAM -g mm -n RYBP &

每个比较都会得到四个文件，如下

NAMEpeaks.xls: 以表格形式存放peak信息，虽然后缀是xls，但其实能用文本编辑器打开，和bed格式类似，但是以1为基，而bed文件是以0为基.也就是说xls的坐标都要减一才是bed文件的坐标
NAMEpeaks.narrowPeak NAMEpeaks.broadPeak 类似。后面4列表示为， integer score for display， fold-change，-log10pvalue，-log10qvalue，relative summit position to peak start。内容和NAMEpeaks.xls基本一致，适合用于导入R进行分析。
NAMEsummits.bed：记录每个peak的peak summits，也就是记录极值点的位置。MACS建议用该文件寻找结合位点的motif。
NAME_model.r，能通过NAME_model.r作图，得到是基于你提供数据的peak模型

• 注：RYBP无peak数据(RYBP_model.r可以用)，手动下载peaks数据

wget ftp://ftp.ncbi.nlm.nih.gov/geo/series/GSE42nnn/GSE42466/suppl/GSE42466_RYBP_peaks_5.txt.gz
gzip -d GSE42466_RYBP_peaks_5.txt.gz
mv GSE42466_RYBP_peaks_5.txt RYBP_summits.bed

6 对peaks注释和可视化(R):Chipseeker包

6.0 安装并加载相应的R包，读入peak文件

source ("https://bioconductor.org/biocLite.R")
biocLite("ChIPseeker")
biocLite("TxDb.Mmusculus.UCSC.mm10.knownGene")
library("ChIPseeker")
library("org.Mm.eg.db")
library("TxDb.Mmusculus.UCSC.mm10.knownGene")
library("clusterProfiler")
txdb <- TxDb.Mmusculus.UCSC.mm10.knownGene

读入peak文件

setwd("F:/Data/chip_seq/aligned")
suz12<-readPeakFile("suz12_peaks.narrowPeak")

6.1 查看peak在全基因组的位置

covplot函数可以计算peak 在染色体上的覆盖区域，并可视化。

covplot(suz12,weightCol=5)

备注：
使用Y叔CHIPseeker包说明书中的covplot(peak, weightCol="V5")提示一下错误：

Error in .normarg_shift_or_weight(weight, "weight", x) : 
  'weight' must be a numeric vector, a single string, or a list-like object

我直接把v去掉了，显示结果和说明书中一致。

suz12 peaks over chromosomes.jpeg

还可以定义具体的染色体(图就不展示了）

covplot(suz12, weightCol=5, chrs=c("chr4", "chr5"), xlim=c(4.5e7, 5e7))

6.2 ChIP peaks结合TSS 区域的情况

• TSS:transcription start site
• 首先，计算ChIP peaks结合在TSS区域的情况。这就需要准备TSS区域，这一般定义在TSS位点的侧翼序列（默认-3000~+3000）。然后比对 map到这些区域的peak，并生成tagMatrix。

promoter <- getPromoters(TxDb=txdb, upstream=3000, downstream=3000)
tagMatrix <- getTagMatrix(suz12, windows=promoter)

• 6.2.1 Heatmap of ChIP binding to TSS regions

tagHeatmap(tagMatrix, xlim=c(-3000, 3000), color="red")

suz12_Heatmap of ChIP peaks binding to TSS regions.jpeg

备注，下面这个代码可以一步生成上面这个图

peakHeatmap(suz12, TxDb=txdb, upstream=3000, downstream=3000, color="blue")

>> preparing promoter regions...     2018-08-17 16:54:26 
>> preparing tag matrix...       2018-08-17 16:54:26 
>> generating figure...      2018-08-17 16:54:41 
>> done...           2018-08-17 16:54:4

6.2.2 Average Profile of ChIP peaks binding to TSS region

plotAvgProf(tagMatrix, xlim=c(-3000, 3000),
            conf=0.95,resample = 1000,
            xlab="Genomic Region (5'->3')", ylab = "Read Count Frequency")

Average Profile of ChIP peaks binding to TSS region.jpeg

6.3 peaks注释

• annotatePeak函数进行peaks注释，可以定义TSS（转录起始位点）区域，默认情况下TSS定义为-3kb到+ 3kb。
• annotatePeak的输出是csAnno格式。 ChIPseeker中国的as.GRanges函数将csAnno转换为GRanges格式，as.data.frame将csAnno转换为data.frame，然后通过write.table将其导出到文件。
• TxDb.Hsapiens.UCSC.hg38.knownGene和TxDb.Hsapiens.UCSC.hg19.knownGene分别对应人类基因组hg38和hg19，TxDb.Mmusculus。
• UCSC.mm10.knownGene和TxDb.Mmusculus.UCSC.mm9.knownGene则对应小鼠基因组mm10和mm9。
• 用户还可以通过RmakeTxDbFromBiomart和makeTxDbFromUCSC从UCSC Genome Bioinformatics和BioMart数据库检索准备自己的TxDb对象。然后进行峰值注释。
• 所有的峰值信息都会保存在输出文件中。其中包含peak最近的gene的位置和链的信息，从peak到最近的gene的TSS的距离等。鉴于某些信息可能的重叠，ChIPseeker采取以下优先级别

Promoter
5’ UTR
3’ UTR
Exon
Intron
Downstream
Intergenic

更加具体信息参考作者用户手册。

peakAnno <- annotatePeak(suz12, tssRegion=c(-3000, 3000),
                         TxDb=txdb, annoDb="org.Mm.eg.db")

>> preparing features information...         2018-08-17 17:24:41 
>> identifying nearest features...       2018-08-17 17:24:41 
>> calculating distance from peak to TSS...  2018-08-17 17:24:42 
>> assigning genomic annotation...       2018-08-17 17:24:42 
>> adding gene annotation...             2018-08-17 17:24:51 
'select()' returned 1:many mapping between keys and columns
>> assigning chromosome lengths          2018-08-17 17:24:52 
>> done...                   2018-08-17 17:24:52 

6.3.1 可视化基因组注释

为了注释给定peak在基因组的位置特征信息，可使用annotatePeak函数，这些信息在输出信息的“annotation”列，包括peak是否在TSS,外显子，5'UTR,3'UTR,内含子或间隔区。很多研究人员对这些注释非常感兴趣。用户可以自己定义TSS区域。

有以下几种可视化方式

• A：pie plot

plotAnnoPie(peakAnno)

Genomic Annotation by pieplot.png

----------------------------------------------------------------

• B: bar plot

plotAnnoBar(peakAnno)

Genomic Annotation by barplot.jpeg

--------------------------------------------------------------

• C: venn plot(重叠）

因为一些注释会重叠，如果想查看重叠的全部注释信息，可以使用veenpie函数

vennpie(peakAnno)

Genomic Annotation by vennpie.jpeg

-----------------------------------------------------------

• D: upsetplot（重叠）

upsetplot(peakAnno)

Genomic Annotation by upsetplot.jpeg

-------------------------------------------------------

6.3.2 Visualize distribution of TF-binding loci relative to TSS

peak（TF结合位点）到最近的gene的TSS之间的距离可以有annotatePeak函数进行计算。作者提供了plotDistToTSS函数计算最近基因的TSS上游和下游的结合位点的百分比，并可视化这种分布。

plotDistToTSS(peakAnno,
              title="Distribution of transcription factor-binding loci\nrelative to TSS")

Distribution of Binding Sites.jpeg

6.4 Functional enrichment analysis功能富集分析

因个人工作原因，最近没有更新，现在已经度过转变期，马上开始更新，感谢关注的朋友及发来私信的朋友。

---

我的博客即将搬运同步至腾讯云+社区，邀请大家一同入驻：https://cloud.tencent.com/developer/support-plan?invite_code=2ys0k0l1st2ck