nievergeltlab’s gists

nievergeltlab / loop_cores.sh

Last active July 25, 2017 15:58

Use all cores within a node via a loop E.g. Given N jobs, with K cores, run jobs 1-K, wait for all to finish, then run K+1 - N

	ncommands=22
	nodesize=6
	nodeuse=$(($nodesize ))
	totjobs=$(( ($ncommands + $nodeuse - 1 ) / $nodeuse ))


	for i in $(seq 1 1 $totjobs)
	do
	jstart=$((($i-1)*$nodesize +1))
	jstop=$(($i*$nodesize))

nievergeltlab / correlate_stats.r

Created July 13, 2017 21:04

Correlate and plot test statistics (pvalues) on a -log10 scale. Useful for comparing two similar analyses.

	/home/laramie/CLEANED_DATA/EXTERNAL_DATA/MIRECC6_Duke/duke_gwas_WH/12_15_15/DUKE_WH_model3_w_A2_danerFORMAT

	grep -v NA MIRE_eur_analysis1_mf > MIRE_eur_analysis1_mf_nona

	adam <- read.table('MIRE_eur_analysis1_mf_nona', nr=10000000,stringsAsFactors=F,header=T)
	lara <- read.table('/home/laramie/CLEANED_DATA/EXTERNAL_DATA/MIRECC6_Duke/duke_gwas_WH/12_15_15/DUKE_WH_model3_w_A2_danerFORMAT', nr=8300000,stringsAsFactors=F,header=T)

	lara$p <- 2*pnorm(abs(log(lara$OR)/lara$SE),lower.tail=F)
	rescor <- merge(adam,lara,by="SNP")

nievergeltlab / familycounter.Rscript

Created July 13, 2017 20:58

Print tables of phenotype information from list of .fam files


	args <- commandArgs(trailingOnly = TRUE)

	family_files <- args[1]

	dat <- read.table(family_files)


	for (i in dat$V1)
	{

nievergeltlab / roc_cox.r

Last active June 29, 2021 21:14

How competing Cox models of risk of relapse can be compared using ROC analysis. This analysis uses an implementation of the permutation methods suggested for comparing competing models

	###################################
	### Load R libraries

	#Contains survival analysis function
	library(survival)
	#Functions to build ROC curves from Cox models
	library(risksetROC)


	###################################

nievergeltlab / heritability.r

Last active November 11, 2025 07:08

Convert heritability from observed scale to liability scale

	#K=pop prevalence
	#P=proportion of cases in study
	#hsq=Heritability estimate (on observed scale)
	#bigT = liability threshold
	#tau = density of gaussian
	K=0.0659
	P=0.0659
	h2=0.0365
	zv <- dnorm(qnorm(K))

nievergeltlab / join.sh

Last active June 29, 2021 21:35

Join two files using the join command. Adjust for localization issues that cause improper sorting.

	LC_ALL=C sort -k1b,1 file1.txt > file1.txt.sorted
	LC_ALL=C sort -k1b,1 file2.txt > file2.txt.sorted
	LC_ALL=C join file1.txt.sorted file2.txt.sorted > joined.txt

nievergeltlab / r_padding.r

Created June 2, 2017 19:19

R Margin Padding With a big font, R will cut off the end of the Y axis label text. This puts padding on the margin to give space for the big text.

	pdf('h2_mf_perstudy.pdf',7,7)
	par(mar=c(5, 4, 4, 2) + 0.5)
	plot(dat$case,dat$SE,cex.axis=1.45,cex.lab=1.5,xlab="N Cases", ylab="Standard Error of SNP Based Heritability",main="",cex=1.5,pch=19)

nievergeltlab / plink_merge.sh

Last active June 30, 2021 16:09

Merge long list of PLINK files


	ls ../bgn \| sed 's/.bed//g' \| sed 's/.bim//g' \| sed 's/.fam//g' \| sort -u \| awk '{print "../bgn/"$1".bed","../bgn/"$1".bim","../bgn/"$1".fam"}' > mergelist.txt

	../plink --merge-list mergelist.txt --maf 0.00000001 --geno 0.02 --make-bed --out pgbd_nomono_goodgeno

nievergeltlab / plink_cov_pheno_merge.r

Created May 1, 2017 19:08

Merge PLINK phenotype and covariates in R

	famfile <- read.table('/home/cnieverg/MEG2/qc/pts_meg2_mix_am.fam',header=F,stringsAsFactors=F)
	names(famfile) <- c("FID","IID","F","M","Gender","Pheno")

	mds <- read.table('/home/cnieverg/MEG2/qc/pca/pts_meg2_mix_am-qc-eu1_pca.menv.mds_cov',header=T,stringsAsFactors=F)

	dat <- merge(famfile,mds,by=c("FID","IID"))


	#Number of analyzed subjects of European ancestry
	dim(dat)[1]

nievergeltlab / quanto_power.r

Created May 1, 2017 19:03

Filter QUANTO outputs into power plots Plot the detectable OR (at a set power and alpha) for a given sample size and MAF

	setwd('C:/Users/adam/Desktop/roycepower')
	library(plyr)

	dat0 <- read.csv('me_power.csv', header=T)

	dat <- dat0

	#Fill in the blanks
	for (i in 1:dim(dat)[1])
	{

Adam X. Maihofer nievergeltlab