nievergeltlab

fo

for study in brya ftca gsdc mrsc nhrv nhs2 nss1 nss2 onga pts1 psy2 psy3 
do
 #cat "$study"/infoeur/*.dosage | awk '{if(NR == 1 || ($1 != "CHR" && $6 > 0 && $6 < 1)) print $2,$1,$3,$4,$5,$6,$7}' >  "$study".mafinfo
 cat "$study"/infoeur/*.dosage | awk '{if(NR == 1 || ($1 != "CHR" && $6 > 0 && $6 < 1)) print $2,$4,$5,$6,$7}' | LC_ALL=C sort -k1b,1 >  "$study".mafinfo
done

#Note: Assumes that the SNP is the first column of your gwas results file. If it is not, change "sort k1b,1" 
to the column number where the SNP is. e.g. if SNP is in column 2, change to sort k2b,2 gwas_results.assoc.txt 

nievergeltlab

nAA=76
nAa=1501
naa=7823
n=nAA+nAa+naa
pA=(2*nAA+nAa)/(2*n)
pa=1-pA
obs= c(nAA,nAa,naa)
exp=c(n*pA^2,2*n*pA*pa,n*pa^2)
val=sum(((obs-exp)^2)/exp)
1-pchisq(val,df=1)

nievergeltlab

library(mice)

#Continuous and categorical variables should be defined prior to MI. This will prevent nonsensical imputations from happening - e.g. cohorts are 1,2,3,4, you impute, then suddenly someone is in cohort 1.5

#Note feb 5: Put other surrogate mobility variables in noting imputation

#I prefer to make lists of all continuous and categorical variables going in

contvars <- scan(what="character",sep="/n")
AGE

nievergeltlab

expand.grid(study= c("pgc","hou"), imputation = c("unimputed","imputed"), subjs = c("pgbd","pgbd_va"), pheno = c("enter_M_v2","nonresponder","eventnonresponder"))

nievergeltlab

for commandz in $(ps aux | grep R | awk '{print $2}'); do kill $commandz;done

nievergeltlab

zcat daner_pts_mrsc_mix_am-qc.hg19.ch.fl.gz  | awk '{print $2,$11}'  | grep -v NA >  daner_pts_mrsc_mix_am-qc_short2
awk '{print $2,$12}' mrsc_gemma_pcs > mrsc_gemma_pcs_short2

gzip -d mrsc_gemma_pcs_short2.gz
gzip -d daner_pts_mrsc_mix_am-qc_short2.gz

LC_ALL=C join <(awk '{if (NR==1)$1="SNP", $2="P"; else print}' mrsc_gemma_pcs_short2 | LC_ALL=C sort -k1b,1 ) <(LC_ALL=C sort -k1b,1 daner_pts_mrsc_mix_am-qc_short2) > files_joined
sort -g -k2 files_joined > files_joined2
grep -v P files_joined2 > files_joined3

nievergeltlab

for files in $(ls *.txt)
do
 samplename=$(awk 'NR==12{print $1}' $files)
 echo $files $samplename >> sample_ids.txt
 samplename=""
done

nievergeltlab

#PBS -lnodes=1
#PBS -lwalltime=02:05:00

#!/bin/bash

#Make walltime a function of N?

while getopts i:f:o:p:d:n: option
do
  case "${option}"

nievergeltlab

setwd('C:/Users/adam/Desktop/starrs_cdrisc')
dat <- read.csv('CD_RISC_v2.csv',header=T,stringsAsFactors=F,na.strings=c("NA","#N/A"))

dat <- subset(dat,visit <= 3)

library(plyr)
cdr_max <- ddply(dat, ~studyid,colwise(.fun=max,'CDR_SUM',na.rm=T))

cdr_max <- cdr_max[-which(cdr_max$CDR_SUM == -Inf),]

nievergeltlab

#############################Begin Function#############################

#First we define the function with required parameters and optional parameters with their defaults

ManhattanPlot_AJS_cut <- function(chr, pos, pvalue, snp, genomesig = 0, genomesug = 0,photoname = "Manhattan Plot", outname = "ManhattanPlot_AJS", colors = c("navyblue","gray69"), sigcolor = "darkred", sugcolor = "indianred", ncex = .5, plotsymbol = 16, sugcex = 1, sigcex = 1.5, sigsnpcolor = "red", nonautosome = c(23,24,25,26),xlabel = "Chromosomal Positions",ylabel = "-log10(p-value)", pdf = "TRUE",topsnplabels = "FALSE", pvalue_miss = "NA",highlight_p=5e-8,highlightboundary=50000) {
	
	

	#bind the input data into one table and prune out missing data based on the pvalue_miss option and then 
	#redefines the input variables chr, pos, pvalue, snp to reflect the cleaned data