Editing GSCAN dbGaP (section)

==Stroke==

===Phenotypes===

<syntaxhighlight lang="rsplus">
### Date: Feb 13 2017
### Author: Scott Vrieze

options(stringsAsFactors=F)

### Load in dataset ###

ninds <- read.table("/work/KellerLab/GSCAN/dbGaP/CIDR_StrokeGenetics/PhenoGenotypeFiles/RootStudyConsentSet_phs000615.CIDR_International_StrokeGenetics.v1.p1.c3.GRU-NPU/PhenotypeFiles/phs000615.v1.pht003307.v1.p1.c3.IschemicStroke_Subject_Phenotypes.GRU-NPU.txt.gz",
                    header = T, sep="\t")

### The file reads into R incorrectly because of a weird trailing tab
### character in the data file, so use the below code to shift column
### names to the correct column.
names(ninds)[1] <- "XXX"
ninds$dbGaP_Subject_ID <- row.names(ninds)
ninds$smokingStatus <- NULL
names(ninds) <- c(names(ninds)[2:17], "smokingStatus", "dbGaP_Subject_ID")
    

### subset the only variables needed
pheno <- subset(ninds, select=c("subject_id", "smokingStatus", "age", "gender", "AFFECTION_STATUS"))


###————————————————————###
### SMOKING INITIATION ###
###————————————————————###

### NINDS variable is “smokingStatus”
### Variables are “NEVER”, “FORMER”, “CURRENT”, “NO”, “UNKNOWN”, and blank.
###
### ’CURRENT' defined as cigarette smoking within last 30 days,
### 'FORMER' defined as ### more than 100 cigarettes in one's lifetime
### but no smoking within the last 30 ### days; ’NEVER' defined as
### less than 100 cigarettes smoked in one's lifetime.
### table(pheno$smokingStatus)
###
###        CURRENT  FORMER   NEVER      NO UNKNOWN
###     63     727    1137    2397       4     251

si <- pheno$smokingStatus
si[si == "CURRENT" | si == "FORMER"] <- 2
si[si == "NEVER"   | si == "NO"] <- 1
si[si != "1" & si != "2"] <- NA
si <- as.numeric(si)

### table(si)
###
###    1    2
### 2401 1864


###————————————————————-###
###  SMOKING CESSATION  ###
###————————————————————-###

### NINDS variable is “smokingStatus”
### Variables are “NEVER”, “FORMER”, “CURRENT”, “NO”, “UNKNOWN”, and blank.
###
### ’CURRENT' defined as cigarette smoking within last 30 days,
### 'FORMER' defined as more than 100 cigarettes in one's lifetime but
### no smoking within the last 30 days; ’NEVER' defined as less than
### 100 cigarettes smoked in one's lifetime.
### table(pheno$smokingStatus)
###
###        CURRENT  FORMER   NEVER      NO UNKNOWN
###     63     727    1137    2397       4     251

sc <- pheno$smokingStatus
sc[sc == "CURRENT"] <- 2
sc[sc == "FORMER"] <- 1
sc[sc != "1" & sc != "2"] <- NA
sc <- as.numeric(sc)

### table(pheno$V5)
###
###    1    2
### 1137  727


###——————————###
###  GENDER  ###
###——————————###

### NINDS variable is “gender”
### Variables are “F” and “M”
### table(pheno$gender)
###
###    F    M
### 2627 1952

sex <- pheno$gender
sex[sex == "M"] <- 1
sex[sex == "F"] <- 2
sex <- as.numeric(sex)

###----------------###
### Write to files ###
###----------------###

### This study uses the "subject_id" ID field in the genotype files,
### so use that here, instead of the dbGaP_Subject_ID
phenotypes <- data.frame(fid = pheno$subject_id,
                     iid = pheno$subject_id,
                     patid = "x",
                     matid = "x",
                     sex = sex,
                     si = si,
                     sc = sc)

write.table(phenotypes,
            "/work/KellerLab/vrieze/GSCAN/GWAS/summary_stats_generated_internally/Stroke/GSCAN_STROKE_phenotypes.ped",
            row.names=F,
            quote = F,
            sep="\t")


covariates  <- data.frame(fid = pheno$subject_id,
                    iid = pheno$subject_id,
                    patid = "x",
                    matid = "x",
                    sex = sex,
                    age = pheno$age,
                    age2 = pheno$age^2,
                    AFFECTION_STATUS = pheno$AFFECTION_STATUS)

write.table(covariates,
            "/work/KellerLab/vrieze/GSCAN/GWAS/summary_stats_generated_internally/Stroke/GSCAN_STROKE_covariates.ped",
            row.names=F,
            quote = F,
            sep="\t") ### save table
</syntaxhighlight>