原貼地址：http://www.lxweimin.com/p/3bfb21d24b74
?????????????????? http://www.lxweimin.com/p/4910d7cec5c8
1.讀取自己表達矩陣
# 構建自己的表達矩陣并讀取
> control1 <- read.table("~/disk2/data/rna-seq/matrix/SRR3589959.count", sep="\t", col.names = c("gene_id","control1"))
> control2 <- read.table("~/disk2/data/rna-seq/matrix/SRR3589961.count", sep="\t", col.names = c("gene_id","control2"))
> rep1 <- read.table("~/disk2/data/rna-seq/matrix/SRR3589960.count", sep="\t", col.names = c("gene_id","akap951"))
> rep2 <- read.table("~/disk2/data/rna-seq/matrix/SRR3589962.count", sep="\t",col.names = c("gene_id","akap952"))
> raw_count <- merge(merge(control1, control2,by="gene_id"),merge(rep1,rep2, by="gene_id"))
> raw_count_filt <- raw_count[-48823:-48825,]
> raw_count_filter <- raw_count_filt[-1:-2,]
> ENSEMBL <- gsub("\\.\\d*","", raw_count_filter$gene_id)
> row.names(raw_count_filter) <- ENSEMBL
> raw_count_filter <- raw_count_filter[ ,-1]
這里有兩個問題，第一個就是刪除行列，自己根據情況選擇，理論上把名字差別太大的刪了。
矩陣數據結構
2.構建dds對象
# 這一步很關鍵，要明白condition這里是因子，不是樣本名稱；小鼠數據有對照組和處理組，各兩個重復
> condition <- factor(c(rep("control",2),rep("akap95",2)), levels = c("control","akap95"))# 獲取count數據
> countData <- raw_count_filter[,1:4]
> colData <- data.frame(row.names=colnames(raw_count_filter), condition)
> dds <- DESeqDataSetFromMatrix(countData, colData, design= ~ condition)# 查看一下dds的內容
> head(dds)
adds概要信息
3.DESeq標準化dds
# normalize 數據
> dds2 <- DESeq(dds)# 查看結果的名稱，本次實驗中是"Intercept"，"condition_akap95_vs_control"
> resultsNames(dds2)# 將結果用results()函數來獲取，賦值給res變量
res <- results(dds2)# summary一下，看一下結果的概要信息summary(res)
result結果可以看到一些基本的信息，p值默認小于0.1，上調基因有625個，下調基因有445個。
res的概要信息
4.提取差異分析結果
# 獲取padj（p值經過多重校驗校正后的值）小于0.05，表達倍數取以2為對數后大于1或者小于-1的差異表達基因。
> table(res$padj<0.05)
> res <- res[order(res$padj),]
> diff_gene_deseq2 <-subset(res,padj <0.05& (log2FoldChange >1| log2FoldChange <-1))
> diff_gene_deseq2 <- row.names(diff_gene_deseq2)
resdata<-merge(as.data.frame(res),as.data.frame(counts(dds2,normalize=TRUE)),by="row.names",sort=FALSE)
# 得到csv格式的差異表達分析結果
> write.csv(resdata,file="control_vs_akap95.cvs",row.names = F)
把名字變成上圖，gsub那個不會整，那就導出csv，然后手工刪除在導入進行后續分析
后續就可以轉換基因名 各種折騰了
require(DOSE)
require(clusterProfiler)
ekk <- enrichKEGG(gene=gene,organism="human",pvalueCutoff=0.01)
ego <- enrichGO(gene=gene,OrgDb="org.Hs.eg.db",ont="CC",pvalueCutoff=0.01,readable=TRUE)
write.csv(summary(ekk),"KEGG-enrich.csv",row.names?=F)
write.csv(summary(ego),"GO-enrich.csv",row.names?=F)
ego <- enrichGO(gene????????? = gene,
universe????? = names(geneList),
OrgDb???????? = org.Hs.eg.db,
ont?????????? = "CC",
pAdjustMethod = "BH",
pvalueCutoff? = 0.01,
qvalueCutoff? = 0.05)
ego2 <- enrichGO(gene???????? = gene.df$ENSEMBL,
OrgDb???????? = org.Hs.eg.db,
keytype?????? = 'ENSEMBL',
ont?????????? = "CC",
pAdjustMethod = "BH",
pvalueCutoff? = 0.01,
qvalueCutoff? = 0.05)
ego3 <- enrichGO(gene???????? = gene.df$SYMBOL,
OrgDb???????? = “org.Hs.eg.db”,
keytype?????? = 'SYMBOL',
ont?????????? = "CC",
pAdjustMethod = "BH",
pvalueCutoff? = 0.01,
qvalueCutoff? = 0.05)
ego <- enrichGO(gene????????? = DEG$,
OrgDb???????? = "org.Hs.eg.db",
ont?????????? = "CC",
pAdjustMethod = "BH",
pvalueCutoff? = 0.01,
qvalueCutoff? = 0.05)
ego <- enrichGO(gene=gene,OrgDb="org.Hs.eg.db",ont="CC",pvalueCutoff=0.01,readable=TRUE)
DO<-enrichDO(gene=DEG$Gene.ID, ont = "DO", pvalueCutoff = 0.01, pAdjustMethod = "BH",qvalueCutoff = 0.05)
eg=bitr(geneID = PP$geneID, "ENTREZID", "SYMBOL", "org.Hs.eg.db")
df.id<-bitr(df$SYMBOL, fromType ="SYMBOL", toType ="ENTREZID",OrgDb ="org.Hs.eg.db")
easy.df<-merge(df,df.id,by="SYMBOL",all=F)
sortdf<-easy.df[order(easy.df$foldChange, decreasing =T),]
gene.expr = sortdf$foldChange
names(gene.expr) <- sortdf$ENTREZID
