Report Example

Shilin Zhao
Department of Biostatistics
Vanderbilt University School of Medicine

2020-06-26

Make/Load Data

set.seed(123)
rawData<-matrix(nrow=200,ncol=10,rnorm(200*10))
row.names(rawData)<-paste0("Sample",1:200)
colnames(rawData)<-paste0("Feature",1:10)
rawData<-data.frame(rawData,FeatureYN1=sample(c(rep(0,100),rep(1,100)),200),FeatureYN2=as.character(sample(c(rep(0,100),rep(1,100)),200)),FeatureGroup1=as.character(sample(c(rep("Pre/Group1",100),rep("Post/Group2",100)),200)),FeatureCategory1=sample(c("D","E"),200,replace = TRUE),FeatureCategory2=sample(c("A","B","C","D"),200,replace = TRUE),stringsAsFactors = TRUE)

#make some NAs
rawData[c(30:40),"Feature6"]=NA
rawData[c(35:70),"Feature7"]=NA
rawData[c(1:6),"FeatureYN2"]=NA

Table: Paired Test

Data format: Two columns, Feature6 (Pre) and Feature7 (Post)

tableOut=summaryTable(rawData,varCols=c("FeatureYN2","Feature6","FeatureCategory1"),varColsPaired=c("FeatureYN1","Feature7","FeatureCategory2"),pairedTest = TRUE)
printSummaryTable(tableOut)
N Pre (200) Post (200) Test Statistic

FeatureYN2 (Pre) vs FeatureYN1 (Post)

McNemar’s chi-squared=0.01; P=0.92
0 51% (196) 51% (99) 50% (97)
1 49% (192) 49% (95) 50% (97)

Feature6 (Pre) vs Feature7 (Post)

Feature6 (189); Feature7 (164); Both=159 -0.673 -0.044 0.582 (0.036±1.055) -0.776 -0.072 0.622 (-0.095±0.985) V=6911 ; P=0.344

FeatureCategory1 (Pre) vs FeatureCategory2 (Post)

A 12% (46) 0% (0) 23% (46)
B 18% (72) 0% (0) 36% (72)
C 10% (40) 0% (0) 20% (40)
D 33% (133) 46% (91) 21% (42)
E 27% (109) 55% (109) 0% (0)
For categorical variable, numbers after proportions are counts; For continuous variable, a b c (x±s). a b c represent the lower quartile a, the median b, and the upper quartile c in different categories. x±s represents Mean±SD.
Tests used: McNemar’s chi-squared test for symmetry of rows and columns in a two-dimensional contingency table; Fisher’s exact test for categorical variable; Wilcoxon Signed Rank Test for continuous variable;

Data format: One data column (Feature6) and one group column (FeatureGroup1, define Pre and Post)

tableOut=summaryTable(rawData,groupCol="FeatureGroup1",varCols=c("FeatureYN2","Feature6","FeatureCategory2"),pairedTest = TRUE)
## Will perform paired test, Please confirm Samples were correctly matched/paired by groupCol.
N Post/Group2 (100) Pre/Group1 (100) Test Statistic

FeatureYN2

McNemar’s chi-squared=0; P=1
0 51% (97) 50% (48) 51% (49)
1 49% (95) 50% (48) 49% (47)

Feature6

Post/Group2 (93); Pre/Group1 (96); Both=93 -0.651 0.125 0.916 (0.093±1.112) -0.653 -0.071 0.535 (0.009±0.948) V=2459 ; P=0.296

FeatureCategory2

McNemar’s chi-squared=4.51; P=0.608
A 23% (46) 26% (26) 20% (20)
B 36% (72) 36% (36) 36% (36)
C 20% (40) 18% (18) 22% (22)
D 21% (42) 20% (20) 22% (22)
For categorical variable, numbers after proportions are counts; For continuous variable, a b c (x±s). a b c represent the lower quartile a, the median b, and the upper quartile c in different categories. x±s represents Mean±SD.
Tests used: McNemar’s chi-squared test for symmetry of rows and columns in a two-dimensional contingency table; Fisher’s exact test for categorical variable; Wilcoxon Signed Rank Test for continuous variable;

Same result in regular paired tests to make sure the table is correct

#First table
matrixForTest=table(rawData$FeatureYN2,rawData$FeatureYN1)
mcnemar.test(matrixForTest)
## 
##  McNemar's Chi-squared test with continuity correction
## 
## data:  matrixForTest
## McNemar's chi-squared = 0.010204, df = 1, p-value = 0.9195
#Second table
dataOneGroup1=as.character(rawData[which(rawData$FeatureYN1==0),"FeatureYN2"])
dataOneGroup2=as.character(rawData[which(rawData$FeatureYN1==1),"FeatureYN2"])
matrixForTest=table(dataOneGroup1,dataOneGroup2)
mcnemar.test(matrixForTest)
## 
##  McNemar's Chi-squared test with continuity correction
## 
## data:  matrixForTest
## McNemar's chi-squared = 0, df = 1, p-value = 1

Table: Not Paired Test

Data format: Two columns, Feature6 (Group 1) and Feature7 (Group 2)

tableOut=summaryTable(rawData,varCols=c("FeatureYN2","Feature6","FeatureCategory1"),varColsPaired=c("FeatureYN1","Feature7","FeatureCategory2"))
printSummaryTable(tableOut)
N Group 1 (200) Group 2 (200) Test Statistic

FeatureYN2 (Group 1) vs FeatureYN1 (Group 2)

X-squared=0.01; P=0.917
0 51% (199) 51% (99) 50% (100)
1 49% (195) 49% (95) 50% (100)

Feature6 (Group 1) vs Feature7 (Group 2)

Feature6 (189); Feature7 (164) -0.653 -0.018 0.591 (0.049±1.027) -0.785 -0.091 0.621 (-0.095±0.994) W=16494 ; P=0.298

FeatureCategory1 (Group 1) vs FeatureCategory2 (Group 2)

X-squared=285.05; P<0.001
A 12% (46) 0% (0) 23% (46)
B 18% (72) 0% (0) 36% (72)
C 10% (40) 0% (0) 20% (40)
D 33% (133) 46% (91) 21% (42)
E 27% (109) 55% (109) 0% (0)
For categorical variable, numbers after proportions are counts; For continuous variable, a b c (x±s). a b c represent the lower quartile a, the median b, and the upper quartile c in different categories. x±s represents Mean±SD.
Tests used: Chi-squared test for categorical variable; Non-Paired Wilcoxon Rank Sum Test for continuous variable;

Data format: One data column (Feature6) and one group column (FeatureGroup1, define Group 1 and Group 2)

N Post/Group2 (100) Pre/Group1 (100) Test Statistic

FeatureYN2

X-squared=0.02; P=0.888
0 51% (99) 50% (48) 52% (51)
1 49% (95) 50% (48) 48% (47)

Feature6

Post/Group2 (93); Pre/Group1 (96) -0.651 0.125 0.916 (0.093±1.112) -0.684 -0.071 0.541 (0.007±0.941) W=4771 ; P=0.415

FeatureCategory2

X-squared=1.28; P=0.734
A 23% (46) 26% (26) 20% (20)
B 36% (72) 36% (36) 36% (36)
C 20% (40) 18% (18) 22% (22)
D 21% (42) 20% (20) 22% (22)
For categorical variable, numbers after proportions are counts; For continuous variable, a b c (x±s). a b c represent the lower quartile a, the median b, and the upper quartile c in different categories. x±s represents Mean±SD.
Tests used: Chi-squared test for categorical variable; Non-Paired Wilcoxon Rank Sum Test for continuous variable;

Table: Same result as Not Paired Test in Hmisc package

outVar="FeatureGroup1"
varForTable1=c("FeatureYN2","Feature6","FeatureCategory2")
formulaForTable<-as.formula(paste0(paste(varForTable1, collapse=" + "),"~",outVar))
s <- summaryM(formulaForTable, data=rawData,
                            overall=TRUE, test=TRUE)

html(s, exclude1=FALSE,  what=c('%'),digits=3, prmsd=TRUE)
Descriptive Statistics (N=200).
N
 Post/Group2
N=100 		Pre/Group1
N=100 		Combined
N=200 		Test Statistic
FeatureYN2 : 0 194 50% (48) 52% (51) 51% (99) χ21=0.08, P=0.7761
  1 50% (48) 48% (47) 49% (95)
Feature6 189 -0.65051 0.12503 0.91599  ( 0.09264 ± 1.11226) -0.68399 -0.07066 0.54079  ( 0.00706 ± 0.94135) -0.65277 -0.01798 0.59106  ( 0.04917 ± 1.02714) F1 187=0.67, P=0.4162
FeatureCategory2 : A 200 26% (26) 20% (20) 23% (46) χ23=1.28, P=0.7341
  B 36% (36) 36% (36) 36% (72)
  C 18% (18) 22% (22) 20% (40)
  D 20% (20) 22% (22) 21% (42)
a b c represent the lower quartile a, the median b, and the upper quartile c for continuous variables. x ± s represents X ± 1 SD.   N is the number of non-missing values. Numbers after proportions are frequencies.
Tests used: 1Pearson test; 2Wilcoxon test .

Table: Continuous Group Variable

temp<-summaryTableContinus(rawData,variables=c("Feature1","Feature2","FeatureYN1","FeatureCategory2"),groupVariable="Feature3")
printSummaryTableContinus(temp)
N Variable Test Statistic

Feature1

200 -0.02 S=1357978, P=0.795

Feature2

200 -0.04 S=1382684, P=0.602

FeatureYN1

200 X2=0.98, P=0.321
0 100 -0.611 -0.004 0.581 (-0.01±0.894)
1 100 -0.49 0.173 0.743 (0.073±1.034)

FeatureCategory2

200 X2=6.92, P=0.075
A 46 -0.821 -0.244 0.315 (-0.26±0.977)
B 72 -0.49 0.173 0.675 (0.047±0.94)
C 40 -0.491 0.339 0.875 (0.161±1.029)
D 42 -0.281 0.13 0.711 (0.202±0.89)
One continuous and one categorical variables: a b c (x±s). a b c represent the lower quartile a, the median b, and the upper quartile c for continuous variable in different categories. x±s represents X±SD.
Two continuous variables: a. a represents the rho statistic of Spearman’s correlation analysis.
Tests used: Kruskal-Wallis test for one continuous and one categorical variables; Spearman correlation for two continuous variables.

Table Example: Showing data.frame in a html table

Sepal.Length Sepal.Width Petal.Length
1 5.1 3.5 1.4
2 4.9 3 1.4
3 4.7 3.2 1.3
A test table
Sepal.Length Sepal.Width Petal.Length
5.1 3.5 1.4
4.9 3.0 1.4
4.7 3.2 1.3