ImpulseDE2 是一个处理时序数据的R包，用于查找时序数据的差异基因，它可以用来解决两类问题。
Case-only differential expression analysis tests, whether the expression level of a gene changes over time.
Case-control differential expression analysis tests, whether the expression trajectory of a gene over time differs between samples from a case and samples from a control condition.
1. Case-only
做差异分析的时候有4个主要的参数：
matCountData ： 输入数据，为基因表达的counts数
dfAnnotation ：构建的样品信息
boolCaseCtrl ： 布尔值，是否执行Case-control分析，默认是FALSE
vecConfounders：向量，为需要修正的批次效应信息
生成模拟数据
library(ImpulseDE2)
lsSimulatedData <- simulateDataSetImpulseDE2(
vecTimePointsA = rep(seq(1,8),3),
vecTimePointsB = NULL,
vecBatchesA = NULL,
vecBatchesB = NULL,
scaNConst = 30,
scaNImp = 10,
scaNLin = 10,
scaNSig = 10,
scaMuBatchEffect = NULL,
scaSDBatchEffect = NULL,
dirOutSimulation = NULL)
其中
lsSimulatedData$matObservedCounts 是Counts数据矩阵，可以用自己的数据代替
lsSimulatedData$dfAnnotation 是 数据的注释信息，可以自行构建
格式如下:
## Sample Condition Time Batch
## A_1_Rep1 A_1_Rep1 case 1 B_NULL
## A_2_Rep1 A_2_Rep1 case 2 B_NULL
## A_3_Rep1 A_3_Rep1 case 3 B_NULL
## A_4_Rep1 A_4_Rep1 case 4 B_NULL
## A_5_Rep1 A_5_Rep1 case 5 B_NULL
## A_6_Rep1 A_6_Rep1 case 6 B_NULL
## A_7_Rep1 A_7_Rep1 case 7 B_NULL
## A_8_Rep1 A_8_Rep1 case 8 B_NULL
## A_1_Rep2 A_1_Rep2 case 1 B_NULL
## A_2_Rep2 A_2_Rep2 case 2 B_NULL
## A_3_Rep2 A_3_Rep2 case 3 B_NULL
## A_4_Rep2 A_4_Rep2 case 4 B_NULL
## A_5_Rep2 A_5_Rep2 case 5 B_NULL
## A_6_Rep2 A_6_Rep2 case 6 B_NULL
## A_7_Rep2 A_7_Rep2 case 7 B_NULL
## A_8_Rep2 A_8_Rep2 case 8 B_NULL
## A_1_Rep3 A_1_Rep3 case 1 B_NULL
## A_2_Rep3 A_2_Rep3 case 2 B_NULL
## A_3_Rep3 A_3_Rep3 case 3 B_NULL
## A_4_Rep3 A_4_Rep3 case 4 B_NULL
## A_5_Rep3 A_5_Rep3 case 5 B_NULL
## A_6_Rep3 A_6_Rep3 case 6 B_NULL
## A_7_Rep3 A_7_Rep3 case 7 B_NULL
## A_8_Rep3 A_8_Rep3 case 8 B_NULL
构建方式如下:
data = lsSimulatedData$matObservedCounts 这里请替换成自己的数据
Sample = colnames(data)
Condition = rep("case",24)
Time = rep(seq(1,8),3)
Batch = rep("B_NULL",24)
dfAnnotation=data.frame(Sample,Condition,Time,Batch,row.names = Sample)
运行：
data = as.matrix(data)
objectImpulseDE2 <- runImpulseDE2(
matCountData = data,
dfAnnotation = dfAnnotation,
boolCaseCtrl = FALSE,
vecConfounders = NULL,
scaNProc = 1 )
查看结果
head(objectImpulseDE2$dfImpulseDE2Results)
2.修正批次效应
假设三个实验重复分别来自三个批次，构建注释信息如下：
Sample = colnames(da)
Condition = rep("case",24)
Time = rep(seq(1,8),3)
B1 = rep("B1",8)
B2 = rep("B2",8)
B3 = rep("B3",8)
Batch = c(B1,B2,B3)
dfAnnotation=data.frame(Sample,Condition,Time,Batch,row.names = Sample)
objectImpulseDE2 <- runImpulseDE2(
matCountData = data,
dfAnnotation = dfAnnotation,
boolCaseCtrl = FALSE,
vecConfounders = c("Batch"), ## 这里选择批次信息
scaNProc = 1 )
3. Case-control
用上面类似的方法构建如下注释信息：
## Sample Condition Time Batch
## A_1_Rep1 A_1_Rep1 case 1 B1
## A_2_Rep1 A_2_Rep1 case 2 B1
## A_3_Rep1 A_3_Rep1 case 3 B1
## A_4_Rep1 A_4_Rep1 case 4 B1
## A_5_Rep1 A_5_Rep1 case 5 B1
## A_6_Rep1 A_6_Rep1 case 6 B1
## A_7_Rep1 A_7_Rep1 case 7 B1
## A_8_Rep1 A_8_Rep1 case 8 B1
## A_1_Rep2 A_1_Rep2 case 1 B2
## A_2_Rep2 A_2_Rep2 case 2 B2
## A_3_Rep2 A_3_Rep2 case 3 B2
## A_4_Rep2 A_4_Rep2 case 4 B2
## A_5_Rep2 A_5_Rep2 case 5 B2
## A_6_Rep2 A_6_Rep2 case 6 B2
## A_7_Rep2 A_7_Rep2 case 7 B2
## A_8_Rep2 A_8_Rep2 case 8 B2
## A_1_Rep3 A_1_Rep3 case 1 B3
## A_2_Rep3 A_2_Rep3 case 2 B3
## A_3_Rep3 A_3_Rep3 case 3 B3
## A_4_Rep3 A_4_Rep3 case 4 B3
## A_5_Rep3 A_5_Rep3 case 5 B3
## A_6_Rep3 A_6_Rep3 case 6 B3
## A_7_Rep3 A_7_Rep3 case 7 B3
## A_8_Rep3 A_8_Rep3 case 8 B3
## B_1_Rep1 B_1_Rep1 control 1 C1
## B_2_Rep1 B_2_Rep1 control 2 C1
## B_3_Rep1 B_3_Rep1 control 3 C1
## B_4_Rep1 B_4_Rep1 control 4 C1
## B_5_Rep1 B_5_Rep1 control 5 C1
## B_6_Rep1 B_6_Rep1 control 6 C1
## B_7_Rep1 B_7_Rep1 control 7 C1
## B_8_Rep1 B_8_Rep1 control 8 C1
## B_1_Rep2 B_1_Rep2 control 1 C2
## B_2_Rep2 B_2_Rep2 control 2 C2
## B_3_Rep2 B_3_Rep2 control 3 C2
## B_4_Rep2 B_4_Rep2 control 4 C2
## B_5_Rep2 B_5_Rep2 control 5 C2
## B_6_Rep2 B_6_Rep2 control 6 C2
## B_7_Rep2 B_7_Rep2 control 7 C2
## B_8_Rep2 B_8_Rep2 control 8 C2
## B_1_Rep3 B_1_Rep3 control 1 C3
## B_2_Rep3 B_2_Rep3 control 2 C3
## B_3_Rep3 B_3_Rep3 control 3 C3
## B_4_Rep3 B_4_Rep3 control 4 C3
## B_5_Rep3 B_5_Rep3 control 5 C3
## B_6_Rep3 B_6_Rep3 control 6 C3
## B_7_Rep3 B_7_Rep3 control 7 C3
## B_8_Rep3 B_8_Rep3 control 8 C3
运行代码进行计算
objectImpulseDE2 <- runImpulseDE2(
matCountData = data,
dfAnnotation = dfAnnotation,
boolCaseCtrl = TRUE,
vecConfounders = c("Batch"),
scaNProc = 1 )
4. 绘制单基因轨迹图
library(ggplot2)
lsgplotsGenes <- plotGenes(
vecGeneIDs = head(result$Gene), ###给定需要绘制的基因列表
scaNTopIDs = NULL, ##绘制最显著的10 个基因
objectImpulseDE2 = objectImpulseDE2,
boolCaseCtrl = T,
dirOut = NULL,
strFileName = NULL,
vecRefPval = NULL,
strNameRefMethod = NULL)
print(lsgplotsGenes[[2]])
5. 绘制表达热图
library(ComplexHeatmap)
lsHeatmaps <- plotHeatmap(
objectImpulseDE2 = objectImpulseDE2,
strCondition = "case",
boolIdentifyTransients = F,
scaQThres = 0.01)
draw(lsHeatmaps$complexHeatmapRaw)
网友评论