Intern

This sofware is for internal use. If you found this page den congrats. Feel free you use the methods but they might not work, give wrong results or other nasty stuff. If and then they fuck up don't contact us or complain.

Secret software

Grepper

Formerly known as getliner. Used for grep'ing with in columns

Sortdep

Fast counting of integers

Other stuff

• chisquare

A small library build using numerical recipies third edition for calculating stuff relating to the chisqdistribution

• bambi/bammer

(Thorfinn, april20 2012)

• • Download

bambi.tar.gz

• • Usage

Program is invoked with either: <example> ./bammer view #samtools view/ single sample only ./bammer uppile #This does emulates the samtools mpileup </example>

Inputfiles are supplied with either <example> tmp.bam #single bamfile -b bam.list #filelist containing bamfiles -b bam.list -nInd INT #only select the first INT samples from bam.list </example>

Regions of interest can be supplied by defining a region in the following way <example> -r chr:a-b #select chromsome chr from a, to b -r chr:-b #select chromosome chr from beginning to be -r chr:b- #select chromosome chr from b to end of chromosome </example>

If we want to dump different regions in the same run, this can be done with <example> -rf regions.file #every region is on a newline, has same format as above </example>

Program can also estimate genotypelikelihoods if uppile is selected

<example> -GL 0 #SOAPsnp -GL 1 #samtools -GL 2 #gatk </example>

Output is hardcode dumped in glfs.glfs, these is a binary double file. Information for first site is the first 10*sizeof(double)*nInd bytes. This is the samtools ordering AA,AC,AG,...

Information about the position is dumped in mafs.mafs. This is NOT the MAF, but simply a 2 column file with chr tab.

Below are some examples for running the program <example>

1. view commands below

./bammer view new2.bam ./bammer view bwape_sort_rmdup.bam chr21:9483252-9672320 ./bammer view bwape_sort_rmdup.bam chrY:-1000000

1. mpileup below

./bammer uppile -b lucampHighBam.filelist -nInd 2 -r chr1:-376920

1. samtools estimation of GL

./bammer uppile -b lucampHighBam.filelist -nInd 2 -r chr1:-376920 -GL 1 </example>

• • Technics

Program spawns a single thread, when a chunk has been read, and the thread is not finished it will wait.

• ieatgor

Extract lines from file in "gor" format. The files can be found in this [[ieatgor/][folder]]

to compile <example> g++ -O3 -o ieatgor ieatgor.cpp -lz </example>

The target file must be sorted. example of target file <example> chr1:22323:232322 chr10:10000:10000000 chr3:10000:10000000

</example>

A "gor" file is a file with the first coloum being the chromosome and the second coloum being the positions. The rest of the coloums are not important. The gor file must be sorted based on the positions and chr. gz gor files are also accepted. example of a gor file

<example>

chr1 1033267 2 3 know 0.000000 EM 0.000000 unk 0.000000 EMunk 0.000000 Nind 9 chr1 1033268 0 2 know 0.000000 EM 0.000000 unk 0.000000 EMunk 0.000000 Nind 9 chr1 1033279 3 1 know 0.000000 EM 0.000000 unk 0.000000 EMunk 0.000000 Nind 9 chr1 1033286 0 2 know 0.000000 EM 0.000000 unk 0.000000 EMunk 0.000000 Nind 9 chr1 1033297 0 2 know 0.000000 EM 0.000000 unk 0.000000 EMunk 0.000000 Nind 9 chr1 1033304 0 1 know 0.000000 EM 0.000000 unk 0.000000 EMunk 0.000000 Nind 9 ...

</example>

run example <example> ./ieatgor file.target file.gor </example> the extracted line are returned to standard out.

• getliner

NB HUGE BUG, must +1 when using -l and zipped files. Sorry for inconveincace. (tsk)

This is a general tool for extract lines from newline seperated textfiles (can be gz compressed). Getliners4 allows for complement grep, use '-v 1'. Default is '-v 0'. Program can also print out basic statistics for hit/nonhit if -i filename is supplied. getliners6.cpp tokinzes on \r, so should work on windows files.

Download here getliners6.cpp . <example> g++ getliners5.cpp -lz -O3 -o getliners </example>

Program can either extract specific lines, or "keys". The linenumbers to extract must be 1-indexed (first line i 1). When using keys, you must supply which column to 'grep' for. The program builds an associative array of the keys. And therefore only a single pass of the datafile is required.

Fields in the datafile can be seperated by a delimiter that can be specified at runtine with the "-d " argument. Default is space/tab. If delimter is an escape sequence it will not work and the code should be modified.

• calcABBABABA

This is program that calculates ABBA-BABA related stats for a given chromosome.

It can be downloaded here abbababatest.cpp and can be compiled with: <example> g++ -O3 abbababatest.cpp -o calcABBABABA -lz </example> To get all options just type <example>

./calcABBABABA 

</example> An example run cmd <example> ../calcABBABABA -file /space/genomes/refgenomes/ancestal/hg19/shortForm/chr20.ans -bam testbamfilelistbotoV2chr20.txt -outfileprefix testbotoV2Chr20plus1x2 -blocksz 5 </example> The program outputs three files: an abbababa file (with only abbababa stats), a count file (with overall counts for the chromosome) and a summary file where the compile time, run time and the cmds and the input files used when the program was run are all listed.

• plot Plink

[[Rfun/][Folder]]

<example>

Rscript /home/software/public/software/intern/Rfun/plink.plot.R <plink output file>

</example>

• ngsAdmix

• • • download

[[ngsAdmix/][Download folder]]

Download path <example> /home/software/public/software/intern/ngsAdmix/ </example>

• • • compile

compile with icpc or g++ (icpc much better) <example> icpc -O3 -o ngsAdmix ngsAdmix12.cpp -lz -pthread g++ -O3 -o ngsAdmix ngsAdmix12.cpp -lz -pthread </example>

• • • options

options are shown with <example> ./ngsAdmix </example> All options also have a short form (first letter)

- default stopping criteria is either tolerence 0.0001 for emSq
- default stopping criteria is difference in loglikelihoods of 0.1 over 50 iterations for both em and emSq
- default optimaziation is emSq

• • • run example

<example> ./ngsAdmix -l dat.beagle -i 5000 -K 3 -P 5 -m 1 -o datK3 </example>

• liftover a plink file

You might have the change the chainFile path and the liftover path if you are not on pontus Note that if two SNP map to the same position after liftover one of them is removed. <example>

1. !/bin/bash
2. make liftover file

Files=$1 liftoverProg=/home/albrecht/bin/prog/liftover/liftOver chainFile=/home/albrecht/bin/prog/liftover/chainFiles/hg18ToHg19.over.chain plink=/home/albrecht/bin/prog/plink-1.07-x86_64/plink

if [ "$#" -eq 0 ]; then

   echo "supply a plink prefix"
   exit

fi

if [ -f $Files.map ] then

   echo using file $Files.map

else

   echo the file ${Files}.map does not exists \(use plink --recode\)

fi

if [ -f $Files.bed ] then

   echo using file $Files.bed

else

   echo the file $Files.bed does not exists  \(use plink --make-bed\)

fi

1. make bed file for liftover

Rscript -e "options(scipen=10);map<-read.table('$Files.map',hea=F,as.is=T);res<-cbind(paste('chr',map[,1],sep=),map[,4]-1,map[,4],map[,2],0,'+');write.table(res,file='$Files.bedfile',col=F,row=F,qu=F,sep='\t')"

1. liftover

$liftoverProg $Files.bedfile $chainFile ${Files}Hg19.bedfile ${Files}Hg19.unmapped

1. rm unmapped

grep -v \# ${Files}Hg19.unmapped | cut -f4 > rmSNPs.txt

1. rm duplicates

cat ${Files}Hg19.bedfile | sort -k1 | rev | uniq -d -f3 | rev | cut -f4 > dubSNP.txt echo removing `wc -l dubSNP.txt` cat dubSNP.txt >> rmSNPs.txt

$plink --noweb --bfile $Files --exclude rmSNPs.txt --recode --out ${Files}Red

1. change pos

Rscript -e "liftmap<-read.table('${Files}Hg19.bedfile',hea=F,colC=c('character','integer')[c(1,2,2,1,2,1)]);liftmap<-read.table('${Files}Hg19.bedfile',hea=F,as.is=T);map<-read.table('${Files}Red.map',hea=F,as.is=T); liftmap<-liftmap[liftmap[,4]%in%map[,2],];rownames(map)<-map[,2];map[liftmap[,4],4]<-liftmap[,3];map[liftmap[,4],1]<-sub('chr',,liftmap[,1]);write.table(map,file='hg19.map',col=F,row=F,qu=F)"

echo The new plink file is called ${Files}Hg19 and are on the same strand as the original file

1. flip strands

cat ${Files}Hg19.bedfile | grep \- | cut -f4 > flipSNPs.txt $plink --noweb --ped ${Files}Red.ped --map hg19.map --flip flipSNPs.txt --recode --make-bed --out ${Files}Hg19 $plink --noweb --bfile ${Files}Hg19 --recode --out ${Files}Hg19 </example>

• • • run example

<example> ./liftover.sh myPlinkFile </example>

• angsd

[[angsd/][Download folder]]

• plot admix results

• • • Download

[[Rfun/plotAdmix_anders.R][Get the R script]]

or the more fancy one

[[Rfun/plotAdmix2.R][fancier plotter (ida)]]

to get the options type

<example> Rscript plotAdmix_anders.R Rscript plotAdmix2.R </example>

• plot LD region from plink file

plot the LD in a region. You need SNPmatrix working. If you cannot get SNPmatrix to work use Relate instead (ldsnp3 function) or if on pontus use

<example> library(snpMatrix,lib="/home/albrecht/R/x86_64-pc-linux-gnu-library/2.15/") </example>

R function: <include file="andersLink/ldplot.R" markup="example">

• • • example

<example> library(snpMatrix)

1. snp info from plink files

pl<-read.plink("/space/anders/greenland/data/plink/datGRDKplus") plB<-read.table("/space/anders/greenland/data/plink/datGRDKplus.bim")

1. choose region

min.pos<-70e6 max.pos<-81e6 chr=1 minMaf<-0.05 minCallrate<-0.95 depth<-400 #depth of ld calculations (window size)

1. calculatate maf and callrate

info<-col.summary(pl)

1. chose SNPs

table(keep <- plB[,1]==chr&plB[,4]>min.pos&plB[,4]<max.pos & info$MAF>minMaf & !is.na(info$MAF) & info$Call.rate > minCallrate)

1. estimate LD

ldinfo <- ld.snp(pl[,keep],depth=depth)

1. plot LD (D')

ldplot(ldinfo$dprime,pos=plB[keep,4],main="LD (D')")

1. plot R squared

ldplot(ldinfo$rsq2,pos=plB[keep,4],main="LD (r2)")

1. change color palette

colR=colorRampPalette(c("white","grey","black"))(40) ldplot(ldinfo$rsq2,pos=plB[keep,4],main="LD (r2)",colR=colR) </example>

This example produces the plots: [[andersLink/ldpic1.pdf][pdf]]