APAqtls with Leafcutter

Filter dose files

I need to remove non snps and snps with <.05 from the dosage file.

I will first copy all of the dosage files to my direcory instead of changing tonys.

cp *dose.vcf.gz /project2/gilad/briana/YRI_geno_hg19/

I want to write a python script that will read in the files and perform the filters.

I wrote a python script that take in the dose file and a name of an out file. I will write a bash script to wrap this on all of the chrs.

#!/bin/bash


#SBATCH --job-name=filter_dose
#SBATCH --account=pi-yangili1
#SBATCH --time=24:00:00
#SBATCH --output=filter_dose.out
#SBATCH --error=filter_dose.err
#SBATCH --partition=broadwl
#SBATCH --mem=12G
#SBATCH --mail-type=END

module load python

for i in 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 
do 
python filter_vcf.py chr$i.dose.vcf chr$i.dose.filt.vcf
done

Now I can use these for the fastqtl script instead.

I also updated to only use the first 2 pcs as covariates.

Run permuted version

Permutation pass to calculate correctedp-values for molecular phenotypes.

APAqtl_perm_tot.sh

#!/bin/bash


#SBATCH --job-name=APAqtl_perm_tot
#SBATCH --account=pi-yangili1
#SBATCH --time=24:00:00
#SBATCH --output=APAqtl_perm_tot.out
#SBATCH --error=APAqtl_perm_tot.err
#SBATCH --partition=broadwl
#SBATCH --mem=12G
#SBATCH --mail-type=END

for i in 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 
do
/home/brimittleman/software/bin/FastQTL/bin/fastQTL.static --permute 1000 --vcf /project2/gilad/briana/YRI_geno_hg19/chr$i.dose.filt.vcf.gz --cov /project2/gilad/briana/threeprimeseq/data/filt_peak_refGene_cov/filtered_APApeaks_merged_allchrom_refseqGenes_pheno_Total.txt.gz.2PCs --bed /project2/gilad/briana/threeprimeseq/data/filt_peak_refGene_cov/filtered_APApeaks_merged_allchrom_refseqGenes_pheno_Total.txt.gz.qqnorm_chr$i.gz --out /project2/gilad/briana/threeprimeseq/data/perm_APAqtl/filtered_APApeaks_merged_allchrom_refseqGenes_pheno_Total.txt.gz.qqnorm_chr$i.perm.out --chunk 1 1  --window 5e4 --include-samples /project2/gilad/briana/threeprimeseq/data/filt_peak_refGene_cov/SAMPLE.txt
done

APAqtl_perm_nuc.sh

#!/bin/bash


#SBATCH --job-name=APAqtl_nominal_nuc
#SBATCH --account=pi-yangili1
#SBATCH --time=24:00:00
#SBATCH --output=APAqtl_perm_nuc.out
#SBATCH --error=APAqtl_perm_nuc.err
#SBATCH --partition=broadwl
#SBATCH --mem=12G
#SBATCH --mail-type=END

for i in 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 
do
/home/brimittleman/software/bin/FastQTL/bin/fastQTL.static --permute 1000 --vcf /project2/gilad/briana/YRI_geno_hg19/chr$i.dose.filt.vcf.gz --cov /project2/gilad/briana/threeprimeseq/data/filt_peak_refGene_cov/filtered_APApeaks_merged_allchrom_refseqGenes_pheno_Nuclear.txt.gz.2PCs --bed /project2/gilad/briana/threeprimeseq/data/filt_peak_refGene_cov/filtered_APApeaks_merged_allchrom_refseqGenes_pheno_Nuclear.txt.gz.qqnorm_chr$i.gz --out /project2/gilad/briana/threeprimeseq/data/perm_APAqtl/filtered_APApeaks_merged_allchrom_refseqGenes_pheno_Nuclear.txt.gz.qqnorm_chr$i.perm.out --chunk 1 1  --window 5e4 --include-samples /project2/gilad/briana/threeprimeseq/data/filt_peak_refGene_cov/SAMPLE.txt
done

The results file has the folowing columns:

• ID of the tested molecular phenotype (in this particular case, the gene ID)
  
• Number of variants tested in cis for this phenotype
  
• MLE of the shape1 parameter of the Beta distribution
  
• MLE of the shape2 parameter of the Beta distribution
  
• Dummy [To be described later]
  
• ID of the best variant found for this molecular phenotypes (i.e. with the smallest p-value)
  
• Distance between the molecular phenotype - variant pair
  
• The nominal p-value of association that quantifies how significant from 0, the regression coefficient is
  
• The slope associated with the nominal p-value of association [only in version > v2-184]
  
• A first permutation p-value directly obtained from the permutations with the direct method. This is basically a corrected version of the nominal p-value that accounts for the fact that multiple variants are tested per molecular phenotype.
  
• A second permutation p-value obtained via beta approximation. We advice to use this one in any downstream analysis.

I can check the experiments as recomended by the FastQTL site.

d = read.table("permutations.all.chunks.txt.gz", hea=F, stringsAsFactors=F)
colnames(d) = c("pid", "nvar", "shape1", "shape2", "dummy", "sid", "dist", "npval", "ppval", "bpval")
plot(d$ppval, d$bpval, xlab="Direct method", ylab="Beta approximation", main="Check plot")
abline(0, 1, col="red")

I will try this first on the resutls from chr1.

nuc.chr1= read.table("../data/perm_QTL/filtered_APApeaks_merged_allchrom_refseqGenes_pheno_Nuclear.txt.gz.qqnorm_chr1.perm.out",head=F, stringsAsFactors=F, col.names = c("pid", "nvar", "shape1", "shape2", "dummy", "sid", "dist", "npval", "slope", "ppval", "bpval"))


plot(nuc.chr1$ppval, nuc.chr1$bpval, xlab="Direct method", ylab="Beta approximation", main="Nuclear Check plot")
abline(0, 1, col="red")

tot.chr1=read.table("../data/perm_QTL/filtered_APApeaks_merged_allchrom_refseqGenes_pheno_Total.txt.gz.qqnorm_chr1.perm.out", head=F, stringsAsFactors = F, col.names= c("pid", "nvar", "shape1", "shape2", "dummy", "sid", "dist", "npval", "slope", "ppval", "bpval"))


plot(tot.chr1$ppval, tot.chr1$bpval, xlab="Direct method", ylab="Beta approximation", main="Total Check plot")
abline(0, 1, col="red")

Correct for multiple testing:

• Bonferonni

nuc.chr1$bonferroni = p.adjust(nuc.chr1$bpval, method="bonferroni")

plot(-log10(nuc.chr1$bonferroni), main="Nuclear chr1 bonferroni corrected pval")

tot.chr1$bonferroni = p.adjust(tot.chr1$bpval, method="bonferroni")

plot(-log10(tot.chr1$bonferroni),  main="Total chr1 bonferroni corrected pval")

< .05 is 1.3 on this plot.

• BH

nuc.chr1$bh=p.adjust(nuc.chr1$bpval, method="fdr")

plot(-log10(nuc.chr1$bh), main="Nuclear chr1 BH corrected pval")

tot.chr1$bh=p.adjust(tot.chr1$bpval, method="fdr")
plot(-log10(tot.chr1$bh), main="Total chr1 BH corrected pval")

10% FDR is 1 on this plot.

Extend to all results:

nuc.res= read.table("../data/perm_QTL/filtered_APApeaks_merged_allchrom_refseqGenes_pheno_Nuclear_permQTLresults.out",head=F, stringsAsFactors=F, col.names = c("pid", "nvar", "shape1", "shape2", "dummy", "sid", "dist", "npval", "slope", "ppval", "bpval"))


plot(nuc.res$ppval, nuc.res$bpval, xlab="Direct method", ylab="Beta approximation", main="Nuclear Check plot")
abline(0, 1, col="red")

tot.res=read.table("../data/perm_QTL/filtered_APApeaks_merged_allchrom_refseqGenes_pheno_Total_permQTLresults.out", head=F, stringsAsFactors = F, col.names= c("pid", "nvar", "shape1", "shape2", "dummy", "sid", "dist", "npval", "slope", "ppval", "bpval"))


plot(tot.res$ppval, tot.res$bpval, xlab="Direct method", ylab="Beta approximation", main="Total Check plot")
abline(0, 1, col="red")

• BH

nuc.res$bh=p.adjust(nuc.res$bpval, method="fdr")

plot(-log10(nuc.res$bh), main="Nuclear BH corrected pval")
abline(h=1, col="red")

tot.res$bh=p.adjust(tot.res$bpval, method="fdr")
plot(-log10(tot.res$bh), main="Total BH corrected pval")
abline(h=1, col="red")

Session information

sessionInfo()

R version 3.5.1 (2018-07-02)
Platform: x86_64-apple-darwin15.6.0 (64-bit)
Running under: macOS Sierra 10.12.6

Matrix products: default
BLAS: /Library/Frameworks/R.framework/Versions/3.5/Resources/lib/libRblas.0.dylib
LAPACK: /Library/Frameworks/R.framework/Versions/3.5/Resources/lib/libRlapack.dylib

locale:
[1] en_US.UTF-8/en_US.UTF-8/en_US.UTF-8/C/en_US.UTF-8/en_US.UTF-8

attached base packages:
[1] stats     graphics  grDevices utils     datasets  methods   base     

loaded via a namespace (and not attached):
 [1] workflowr_1.1.1   Rcpp_0.12.18      digest_0.6.15    
 [4] rprojroot_1.3-2   R.methodsS3_1.7.1 backports_1.1.2  
 [7] git2r_0.23.0      magrittr_1.5      evaluate_0.11    
[10] stringi_1.2.4     whisker_0.3-2     R.oo_1.22.0      
[13] R.utils_2.6.0     rmarkdown_1.10    tools_3.5.1      
[16] stringr_1.3.1     yaml_2.1.19       compiler_3.5.1   
[19] htmltools_0.3.6   knitr_1.20