ChromHMM analysis

Pull one set of random snps:

I do still need to get 880 random snps.

shuf -n 880 /project2/gilad/briana/threeprimeseq/data/nominal_APAqtl_trans/filtered_APApeaks_merged_allchrom_refseqGenes_pheno_Nuclear_NomRes.txt > /project2/gilad/briana/threeprimeseq/data/nominal_APAqtl_trans/randomSnps/ApaQTL_nuclear_Random880.txt

Run QTLNOMres2SigSNPbed.py with nuclear 880 and sort output

import pybedtools 

RANDnuc=pybedtools.BedTool('/project2/gilad/briana/threeprimeseq/data/nominal_APAqtl_trans/randomSnps/ApaQTL_nuclear_Random880.sort.bed') 



hmm=pybedtools.BedTool("/project2/gilad/briana/genome_anotation_data/GM12878.chromHMM.sort.bed")

#map hmm to snps  
NucRnad_overlapHMM=RANDnuc.map(hmm, c=4)


#save results  

NucRnad_overlapHMM.saveas("/project2/gilad/briana/threeprimeseq/data/nominal_APAqtl_trans/randomSnps/ApaQTL_nuclear_Random_overlapHMM.bed")

NuclearRandOverlapHMM=read.table("../data/ChromHmmOverlap/ApaQTL_nuclear_Random_overlapHMM.bed", col.names=c("chrom", "start", "end", "sid", "significance", "strand", "number"))

NuclearRandOverlapHMM_names=NuclearRandOverlapHMM %>% left_join(chromHmm, by="number")

ggplot(NuclearRandOverlapHMM_names, aes(x=name)) + geom_bar() + labs(title="ChromHMM labels for Nuclear APAQtls (Random)" , y="Number of SNPs", x="Region")+theme(axis.text.x = element_text(angle = 90, hjust = 1))

To put this on the same plot I can count the number in each then plot them next to eachother.

random_perChromHMM_nuc=NuclearRandOverlapHMM_names %>%  group_by(name) %>% summarise(Random=n())
sig_perChromHMM_nuc= NuclearOverlapHMM_names %>%  group_by(name) %>%  summarise(Nuclear_QTLs=n())

perChrommHMM_nuc=random_perChromHMM_nuc %>%  full_join(sig_perChromHMM_nuc, by="name", ) %>% replace_na(list(Random=0,Total_QTLs=0))  

perChrommHMM_nuc_melt=melt(perChrommHMM_nuc, id.vars="name")
names(perChrommHMM_nuc_melt)=c("Region","Set", "N_Snps" )

chromenrichNuclearplot=ggplot(perChrommHMM_nuc_melt, aes(x=Region, y=N_Snps, by=Set, fill=Set)) + geom_bar(position="dodge", stat="identity") +theme(axis.text.x = element_text(angle = 90, hjust = 1)) + labs(title="Enrichment of Nuclear QTLs by chromatin region", y="Number of Snps", x="Chromatin Region") + scale_fill_brewer(palette="Paired")
chromenrichNuclearplot

ggsave("../output/plots/ChromHmmEnrich_Nuclear.png", chromenrichNuclearplot)

Saving 7 x 5 in image

perChrommHMM_nuc$Random= as.integer(perChrommHMM_nuc$Random)
perChrommHMM_nuc_enr=perChrommHMM_nuc %>%  mutate(Nuclear=Nuclear_QTLs-Random)

perChrommHMM_tot_enr=read.table("../data/ChromHmmOverlap/perChrommHMM_Total_enr.txt",stringsAsFactors = F,header = T)

allenrich=perChrommHMM_tot_enr %>% inner_join(perChrommHMM_nuc_enr, by="name") %>% select(name, Total, Nuclear)

allenrich_melt=melt(allenrich, id.vars="name")

chromenrichBoth=ggplot(allenrich_melt, aes(x=name, by=variable, y=value, fill=variable)) + geom_bar(stat="identity", position = "dodge") + theme(axis.text.x = element_text(angle = 90, hjust = 1)) + labs(title="QTL-Random for each bin by fraction", y="Num QTL SNPs - Num Random SNPs") + scale_fill_manual(values=c("darkviolet", "deepskyblue3"))


ggsave("../output/plots/ChromHmmEnrich_BothFrac.png", chromenrichBoth)

Saving 7 x 5 in image

Permutations

I want to permute the background snps so i can get a better expectation. To do this I need to chose random lines from the nominal file, change the lines to snp format, overlap with HMM, count how many are in each category, and append the list to a dataframe that is category by permuation.

DO this for total first (118 snps)

total_random118_chromHmm.sh

#!/bin/bash

#SBATCH --job-name=total_random118_chromHmm_f
#SBATCH --account=pi-yangili1
#SBATCH --time=36:00:00
#SBATCH --output=total_random118_chromHmm_f.out
#SBATCH --error=total_random118_chromHmm_f.err
#SBATCH --partition=bigmem2
#SBATCH --mem=200G
#SBATCH --mail-type=END

module load Anaconda3
source activate three-prime-env


#test with 2 permutations then make it 1000  
#choose random res
for i in {1..1000};
do
shuf -n 118 /project2/gilad/briana/threeprimeseq/data/nominal_APAqtl_trans/filtered_APApeaks_merged_allchrom_refseqGenes_pheno_Total_NomRes.txt > /project2/gilad/briana/threeprimeseq/data/random_QTLsnps/Total/randomRes_Total_118_${i}.txt
done

#make random 
for i in {1..1000};
do
python randomRes2SNPbed.py Total 118 ${i}
done 


#cat res together   
cat /project2/gilad/briana/threeprimeseq/data/random_QTLsnps/Total/snp_bed/* > /project2/gilad/briana/threeprimeseq/data/random_QTLsnps/Total/snp_bed_all/randomRes_Total_118_ALLperm.bed


#sort full file 
sort -k1,1 -k2,2n /project2/gilad/briana/threeprimeseq/data/random_QTLsnps/Total/snp_bed_all/randomRes_Total_118_ALLperm.bed > /project2/gilad/briana/threeprimeseq/data/random_QTLsnps/Total/snp_bed_all/randomRes_Total_118_ALLperm.sort.bed


#hmm overlap
python overlap_chromHMM.py  Total 118 1000

#Next I would pull this into R to do the group by and average!

pull_random_lines.py

def main(inFile, outFile ,nsamp):
  nom_res= pd.read_csv(inFile, sep="\t", encoding="utf-8",header=None)
  out=open(outFile, "w")
  sample=nom_res.sample(nsamp)
  sample.to_csv(out, sep="\t", encoding='utf-8', index=False, header=F)
  out.close()
    
if __name__ == "__main__":
    import sys
    import pandas as pd
    fraction = sys.argv[1]
    nsamp=sys.argv[2]
    nsamp=int(nsamp)
    iter=sys.argv[3]
    inFile = "/project2/gilad/briana/threeprimeseq/data/nominal_APAqtl_trans/filtered_APApeaks_merged_allchrom_refseqGenes_pheno_%s_NomRes.txt"%(fraction)
    outFile = "/project2/gilad/briana/threeprimeseq/data/random_QTLsnps/%s/randomRes_%s_%d_%s.txt"%(fraction,fraction, nsamp, iter)
    main(inFile, outFile, nsamp)

randomRes2SNPbed.py

def main(inFile, outFile):
    fout=open(outFile, "w")
    fin=open(inFile, "r")
    for ln in fin:
          pid, sid, dist, pval, slope = ln.split()
          chrom, pos= sid.split(":")
          name=sid
          start= int(pos)-1
          end=int(pos)
          strand=pid.split(":")[3].split("_")[1]
          pval=float(pval)
          fout.write("%s\t%s\t%s\t%s\t%s\t%s\n"%(chrom, start, end, name, pval, strand))
    fout.close()

if __name__ == "__main__":
    import sys
    fraction=sys.argv[1]
    nsamp=sys.argv[2]
    nsamp=int(nsamp)
    iter=sys.argv[3]
    inFile = "/project2/gilad/briana/threeprimeseq/data/random_QTLsnps/%s/randomRes_%s_%d_%s.txt"%(fraction,fraction, nsamp, iter)
    outFile= "/project2/gilad/briana/threeprimeseq/data/random_QTLsnps/%s/snp_bed/randomRes_%s_%d_%s.bed"%(fraction,fraction, nsamp, iter)
    main(inFile,outFile) 

overlap_chromHMM.py

def main(inFile, outFile):
  rand=pybedtools.BedTool(inFile) 
  hmm=pybedtools.BedTool("/project2/gilad/briana/genome_anotation_data/GM12878.chromHMM.sort.bed")
  #map hmm to snps
  Rand_overlapHMM=rand.map(hmm, c=4)
  #save results
  Rand_overlapHMM.saveas(outFile)


if __name__ == "__main__":
    import sys
    import pandas as pd
    import pybedtools
    fraction=sys.argv[1]
    nsamp=sys.argv[2]
    niter=sys.argv[3]
    inFile = "/project2/gilad/briana/threeprimeseq/data/random_QTLsnps/%s/snp_bed_all/randomRes_%s_%s_ALLperm.sort.bed"%(fraction,fraction, nsamp)
    outFile= "/project2/gilad/briana/threeprimeseq/data/random_QTLsnps/%s/chromHMM_overlap/randomres_overlapChromHMM_%s_%s_%s.txt"%(fraction,fraction,nsamp, niter)
    main(inFile,outFile)

*Nuclear 880

nuclear_random880_chromHmm.sh

#!/bin/bash

#SBATCH --job-name=nuc_random880_chromHmm
#SBATCH --account=pi-yangili1
#SBATCH --time=36:00:00
#SBATCH --output=nuc_random880_chromHmm.out
#SBATCH --error=nuc_random880_chromHmm.err
#SBATCH --partition=bigmem2
#SBATCH --mem=200G
#SBATCH --mail-type=END

module load Anaconda3
source activate three-prime-env


#test with 2 permutations then make it 1000  
#choose random res
for i in {1..1000};
do
shuf -n 880 /project2/gilad/briana/threeprimeseq/data/nominal_APAqtl_trans/filtered_APApeaks_merged_allchrom_refseqGenes_pheno_Nuclear_NomRes.txt > /project2/gilad/briana/threeprimeseq/data/random_QTLsnps/Nuclear/randomRes_Nuclear_880_${i}.txt
done

#make random 
for i in {1..1000};
do
python randomRes2SNPbed.py Nuclear 880 ${i} 
done 


#cat res together   
cat /project2/gilad/briana/threeprimeseq/data/random_QTLsnps/Nuclear/snp_bed/* > /project2/gilad/briana/threeprimeseq/data/random_QTLsnps/Nuclear/snp_bed_all/randomRes_Nuclear_880_ALLperm.bed


#sort full file 
sort -k1,1 -k2,2n /project2/gilad/briana/threeprimeseq/data/random_QTLsnps/Nuclear/snp_bed_all/randomRes_Nuclear_880_ALLperm.bed > /project2/gilad/briana/threeprimeseq/data/random_QTLsnps/Nuclear/snp_bed_all/randomRes_Nuclear_880_ALLperm.sort.bed


#hmm overlap
python overlap_chromHMM.py  Nuclear 880 1000

#Next I would pull this into R to do the group by and average!

Perm didnt finish: do this with less (824)

nuclear_random880_chromHmm.sm.sh

#!/bin/bash

#SBATCH --job-name=nuc_random880_chromHmm_sm
#SBATCH --account=pi-yangili1
#SBATCH --time=24:00:00
#SBATCH --output=nuc_random880_chromHmm_sm.out
#SBATCH --error=nuc_random880_chromHmm_sm.err
#SBATCH --partition=bigmem2
#SBATCH --mem=100G
#SBATCH --mail-type=END

module load Anaconda3
source activate three-prime-env
#make random 
for i in {1..824};
do
python randomRes2SNPbed.py Nuclear 880 ${i} 
done 


#cat res together   
cat /project2/gilad/briana/threeprimeseq/data/random_QTLsnps/Nuclear/snp_bed/* > /project2/gilad/briana/threeprimeseq/data/random_QTLsnps/Nuclear/snp_bed_all/randomRes_Nuclear_880_ALLperm.bed


#sort full file 
sort -k1,1 -k2,2n /project2/gilad/briana/threeprimeseq/data/random_QTLsnps/Nuclear/snp_bed_all/randomRes_Nuclear_880_ALLperm.bed > /project2/gilad/briana/threeprimeseq/data/random_QTLsnps/Nuclear/snp_bed_all/randomRes_Nuclear_880_ALLperm.sort.bed


#hmm overlap
python overlap_chromHMM.py  Nuclear 880 824

I need a way to make this more efficient to run 1000 permutations. Here I will look at the results from the 824 permutations.

nuclear_perm824= read.table("../data/ChromHmmOverlap/randomres_overlapChromHMM_Nuclear_880_824.txt", col.names=c("chrom", "start", "end", "sid", "significance", "strand", "number"),stringsAsFactors = F, na.strings = "NA")
#924 snps are not annoated 

nuclear_perm824$number=as.integer(as.factor(nuclear_perm824$number))

nuclear_perm824_names=nuclear_perm824 %>% left_join(chromHmm, by="number")

random_perChromHMM_nuc_PERM=nuclear_perm824_names %>%  group_by(name) %>% summarise(Random=n()) %>% mutate(Random_perm=Random/824) %>%  replace_na(list(name="No_annoation")) 

perChrommHMM_nuc_withPerm=random_perChromHMM_nuc_PERM %>%  full_join(sig_perChromHMM_nuc, by="name" ) %>% replace_na(list(Random=0,Nuclear_QTLs=0)) %>%  select(name,Random_perm, Nuclear_QTLs)

 

perChrommHMM_nuc_withPerm_melt=melt(perChrommHMM_nuc_withPerm, id.vars="name")
names(perChrommHMM_nuc_withPerm_melt)=c("Region","Set", "N_Snps" )




ggplot(perChrommHMM_nuc_withPerm_melt, aes(x=Region, y=N_Snps, by=Set, fill=Set)) + geom_bar(position="dodge", stat="identity") +theme(axis.text.x = element_text(angle = 90, hjust = 1)) + labs(title="Enrichment of Nuclear QTLs by chromatin region", y="Number of Snps", x="Chromatin Region") + scale_fill_brewer(palette="Paired")

Enrichment is the actual/random:

perChrommHMM_nuc_withPerm_enrich = perChrommHMM_nuc_withPerm %>% mutate(Nuclear_Enrichment=(Nuclear_QTLs-Random_perm)/Random_perm, chiSq=(Nuclear_QTLs-Random_perm)^2/Random_perm)

ggplot(perChrommHMM_nuc_withPerm_enrich, aes(x=name, y=Nuclear_Enrichment)) + geom_bar(stat="identity",fill="deepskyblue3")+ theme(axis.text.x = element_text(angle = 90, hjust = 1)) + labs(title="ChromHMM Enrichment of Nuclear ApaQTLs \n over 824 Random Permuations", x="Region")

ggplot(perChrommHMM_nuc_withPerm_enrich, aes(x=name, y=chiSq)) + geom_bar(stat="identity",fill="deepskyblue3")+ theme(axis.text.x = element_text(angle = 90, hjust = 1)) + labs(title="ChromHMM ChiSq of Nuclear ApaQTLs \n over 824 Random Permuations", x="Region") 

To parallelize this I will run the permutations in 4 bash scripts:

nuc_random880_chromHmm_set1.sh

#!/bin/bash

#SBATCH --job-name=nuc_random880_chromHmm_set1
#SBATCH --account=pi-yangili1
#SBATCH --time=24:00:00
#SBATCH --output=nuc_random880_chromHmm_set1.out
#SBATCH --error=nuc_random880_chromHmm_set1.err
#SBATCH --partition=bigmem2
#SBATCH --mem=100G
#SBATCH --mail-type=END

module load Anaconda3
source activate three-prime-env
#make random 
for i in {1..250};
do
shuf -n 880 /project2/gilad/briana/threeprimeseq/data/nominal_APAqtl_trans/filtered_APApeaks_merged_allchrom_refseqGenes_pheno_Nuclear_NomRes.txt > /project2/gilad/briana/threeprimeseq/data/random_QTLsnps/Nuclear/randomRes_Nuclear_880_${i}.txt
done

nuc_random880_chromHmm_set2.sh

#!/bin/bash

#SBATCH --job-name=nuc_random880_chromHmm_set2
#SBATCH --account=pi-yangili1
#SBATCH --time=24:00:00
#SBATCH --output=nuc_random880_chromHmm_set2.out
#SBATCH --error=nuc_random880_chromHmm_set2.err
#SBATCH --partition=bigmem2
#SBATCH --mem=200G
#SBATCH --mail-type=END

module load Anaconda3
source activate three-prime-env
#make random 
for i in {251..500};
do
shuf -n 880 /project2/gilad/briana/threeprimeseq/data/nominal_APAqtl_trans/filtered_APApeaks_merged_allchrom_refseqGenes_pheno_Nuclear_NomRes.txt > /project2/gilad/briana/threeprimeseq/data/random_QTLsnps/Nuclear/randomRes_Nuclear_880_${i}.txt
done

nuc_random880_chromHmm_set3.sh

#!/bin/bash

#SBATCH --job-name=nuc_random880_chromHmm_set3
#SBATCH --account=pi-yangili1
#SBATCH --time=24:00:00
#SBATCH --output=nuc_random880_chromHmm_set3.out
#SBATCH --error=nuc_random880_chromHmm_set3.err
#SBATCH --partition=bigmem2
#SBATCH --mem=200G
#SBATCH --mail-type=END

module load Anaconda3
source activate three-prime-env
#make random 
for i in {501..750};
do
shuf -n 880 /project2/gilad/briana/threeprimeseq/data/nominal_APAqtl_trans/filtered_APApeaks_merged_allchrom_refseqGenes_pheno_Nuclear_NomRes.txt > /project2/gilad/briana/threeprimeseq/data/random_QTLsnps/Nuclear/randomRes_Nuclear_880_${i}.txt
done

nuc_random880_chromHmm_set4.sh

#!/bin/bash

#SBATCH --job-name=nuc_random880_chromHmm_set4
#SBATCH --account=pi-yangili1
#SBATCH --time=24:00:00
#SBATCH --output=nuc_random880_chromHmm_set4.out
#SBATCH --error=nuc_random880_chromHmm_set4.err
#SBATCH --partition=bigmem2
#SBATCH --mem=200G
#SBATCH --mail-type=END

module load Anaconda3
source activate three-prime-env
#make random 
for i in {751..1000};
do
shuf -n 880 /project2/gilad/briana/threeprimeseq/data/nominal_APAqtl_trans/filtered_APApeaks_merged_allchrom_refseqGenes_pheno_Nuclear_NomRes.txt > /project2/gilad/briana/threeprimeseq/data/random_QTLsnps/Nuclear/randomRes_Nuclear_880_${i}.txt
done

Same for total:

total_random118_chromHmm_set1.sh

#!/bin/bash

#SBATCH --job-name=total_random118_chromHmm_set1
#SBATCH --account=pi-yangili1
#SBATCH --time=36:00:00
#SBATCH --output=total_random118_chromHmm_set1.out
#SBATCH --error=total_random118_chromHmm_set1.err
#SBATCH --partition=bigmem2
#SBATCH --mem=200G
#SBATCH --mail-type=END

module load Anaconda3
source activate three-prime-env


#test with 2 permutations then make it 1000  
#choose random res
for i in {1..250};
do
shuf -n 118 /project2/gilad/briana/threeprimeseq/data/nominal_APAqtl_trans/filtered_APApeaks_merged_allchrom_refseqGenes_pheno_Total_NomRes.txt > /project2/gilad/briana/threeprimeseq/data/random_QTLsnps/Total/randomRes_Total_118_${i}.txt
done

total_random118_chromHmm_set2.sh

#!/bin/bash

#SBATCH --job-name=total_random118_chromHmm_set2
#SBATCH --account=pi-yangili1
#SBATCH --time=36:00:00
#SBATCH --output=total_random118_chromHmm_set2.out
#SBATCH --error=total_random118_chromHmm_set2.err
#SBATCH --partition=bigmem2
#SBATCH --mem=200G
#SBATCH --mail-type=END

module load Anaconda3
source activate three-prime-env


#test with 2 permutations then make it 1000  
#choose random res
for i in {251..500};
do
shuf -n 118 /project2/gilad/briana/threeprimeseq/data/nominal_APAqtl_trans/filtered_APApeaks_merged_allchrom_refseqGenes_pheno_Total_NomRes.txt > /project2/gilad/briana/threeprimeseq/data/random_QTLsnps/Total/randomRes_Total_118_${i}.txt
done

total_random118_chromHmm_set4.sh

#!/bin/bash

#SBATCH --job-name=total_random118_chromHmm_set4
#SBATCH --account=pi-yangili1
#SBATCH --time=36:00:00
#SBATCH --output=total_random118_chromHmm_set4.out
#SBATCH --error=total_random118_chromHmm_set4.err
#SBATCH --partition=bigmem2
#SBATCH --mem=200G
#SBATCH --mail-type=END

module load Anaconda3
source activate three-prime-env


#test with 2 permutations then make it 1000  
#choose random res
for i in {751..1000};
do
shuf -n 118 /project2/gilad/briana/threeprimeseq/data/nominal_APAqtl_trans/filtered_APApeaks_merged_allchrom_refseqGenes_pheno_Total_NomRes.txt > /project2/gilad/briana/threeprimeseq/data/random_QTLsnps/Total/randomRes_Total_118_${i}.txt
done

I want to turn each of these into snp files:

randomLines2Snp.sh

#!/bin/bash

#SBATCH --job-name=randomLines2Snp
#SBATCH --account=pi-yangili1
#SBATCH --time=36:00:00
#SBATCH --output=randomLines2Snp.out
#SBATCH --error=randomLines2Snp.err
#SBATCH --partition=broadwl
#SBATCH --mem=50G
#SBATCH --mail-type=END

module load Anaconda3
source activate three-prime-env


#make random 
for i in {1..1000};
do
python randomRes2SNPbed.py Nuclear 880 ${i} 
done 

#make random 
for i in {1..1000};
do
python randomRes2SNPbed.py Total 118 ${i}
done 

Next step is the overlap. I want this to run on each seperatly.

sortRandomSnps.sh

#!/bin/bash

#SBATCH --job-name=sortRandomSnps
#SBATCH --account=pi-yangili1
#SBATCH --time=36:00:00
#SBATCH --output=sortRandomSnps.out
#SBATCH --error=sortRandomSnps.err
#SBATCH --partition=broadwl
#SBATCH --mem=50G
#SBATCH --mail-type=END

module load Anaconda3
source activate three-prime-env


for i in $(ls /project2/gilad/briana/threeprimeseq/data/random_QTLsnps/Nuclear/snp_bed/);
do
sort -k1,1 -k2,2n /project2/gilad/briana/threeprimeseq/data/random_QTLsnps/Nuclear/snp_bed/$i > /project2/gilad/briana/threeprimeseq/data/random_QTLsnps/Nuclear/snp_bed_sort/$i.sort.bed
done

for i in $(ls /project2/gilad/briana/threeprimeseq/data/random_QTLsnps/Total/snp_bed/);
do
sort -k1,1 -k2,2n /project2/gilad/briana/threeprimeseq/data/random_QTLsnps/Total/snp_bed/$i > /project2/gilad/briana/threeprimeseq/data/random_QTLsnps/Total/snp_bed_sort/$i.sort.bed
done

Rewrite overlap with ChromHMM script to do it on each file seperatly.

overlap_chromHMM_sepfiles.py

def main(inFile, outFile):
  rand=pybedtools.BedTool(inFile) 
  hmm=pybedtools.BedTool("/project2/gilad/briana/genome_anotation_data/GM12878.chromHMM.sort.bed")
  #map hmm to snps
  Rand_overlapHMM=rand.map(hmm, c=4)
  #save results
  Rand_overlapHMM.saveas(outFile)


if __name__ == "__main__":
    import sys
    import pandas as pd
    import pybedtools
    fraction=sys.argv[1]
    nsamp=sys.argv[2]
    niter=sys.argv[3]
    #which itteration we are on 
    inFile ="/project2/gilad/briana/threeprimeseq/data/random_QTLsnps/%s/snp_bed_sort/randomRes_%s_%s_%s.bed.sort.bed"%(fraction,fraction, nsamp, iter)
    outFile= "/project2/gilad/briana/threeprimeseq/data/random_QTLsnps/%s/chromHMM_overlap/randomres_overlapChromHMM_%s_%s_%s.txt"%(fraction,fraction,nsamp, niter)
    main(inFile,outFile)

overlap_chromHMM_sepfiles.sh

#!/bin/bash

#SBATCH --job-name=overlap_chromHMM_sepfiles
#SBATCH --account=pi-yangili1
#SBATCH --time=36:00:00
#SBATCH --output=overlap_chromHMM_sepfiles.out
#SBATCH --error=overlap_chromHMM_sepfiles.err
#SBATCH --partition=broadwl
#SBATCH --mem=50G
#SBATCH --mail-type=END

module load Anaconda3
source activate three-prime-env

for i in {1..1000};
do
python overlap_chromHMM_sepfiles.py  Nuclear 880 $i
done

for i in {1..1000};
do
python overlap_chromHMM_sepfiles.py  Total 118 $i
done

I will next make an R script that will take in each file and perform the groupby command to get the number of snps in each group.

groupRandomByChromHMM.R

#!/bin/rscripts

# usage: groupRandomByChromHMM.R -f infile -o outfile 

#this file will take any of the itterations and output a file with chrom hmm number, name, numberof snps

library(optparse)
library(dplyr)
library(tidyr)
library(ggplot2)
library(readr)

option_list = list(
  make_option(c("-f", "--file"), action="store", default=NA, type='character',
              help="input coverage file"),
  make_option(c("-o", "--output"), action="store", default=NA, type='character',
              help="output file")
)

opt_parser <- OptionParser(option_list=option_list)
opt <- parse_args(opt_parser)


#interrupt execution if no file is  supplied
if (is.null(opt$file)){
  print_help(opt_parser)
  stop("Need input file", call.=FALSE)
}
if (is.null(opt$output)){
  print_help(opt_parser)
  stop("Need output file", call.=FALSE)
}

randomSNPS=read.table(opt$file, col.names=c("chrom", "start", "end", "sid", "significance", "strand", "number"),stringsAsFactors = F, na.strings = "NA")
hmm_names=read.table("/project2/gilad/briana/genome_anotation_data/chromHMM_regions.txt", col.names = c("number", "name"),stringsAsFactors=F)
randomSNPS$number=as.integer(as.factor(randomSNPS$number))
randomSNPS_names= randomSNPS  %>% left_join(hmm_names, by="number")
#split the name of the file to get the iteration number
fileSplit=strsplit(opt$file, "/")[[1]][10]
iter.txt=strsplit(fileSplit, "_")[[1]][5]
iter=substr(iter.txt, 1, nchar(iter.txt)-4) 

randomSNPS_names_grouped=randomSNPS_names %>%  group_by(number) %>% summarise(!!iter:=n()) %>%  replace_na(list(name="No_annotation")) %>%  dplyr::select(number, !!iter) 
hmm_names$number=as.character(hmm_names$number)
final=hmm_names %>% left_join(randomSNPS_names_grouped,by="number")

write.table(final,opt$output,quote=FALSE, col.names = T, row.names = F)

groupRandomChromHMM.sh

#!/bin/bash

#SBATCH --job-name=groupRandomChromHMM
#SBATCH --account=pi-yangili1
#SBATCH --time=36:00:00
#SBATCH --output=groupRandomChromHMM.out
#SBATCH --error=groupRandomChromHMM.err
#SBATCH --partition=broadwl
#SBATCH --mem=50G
#SBATCH --mail-type=END


module load Anaconda3
source activate three-prime-env

for i in {1..1000};
do
Rscript groupRandomByChromHMM.R -f /project2/gilad/briana/threeprimeseq/data/random_QTLsnps/Nuclear/chromHMM_overlap/randomres_overlapChromHMM_Nuclear_880_${i}.txt -o /project2/gilad/briana/threeprimeseq/data/random_QTLsnps/Nuclear/chromHMM_overlap_group/randomres_overlapChromHMM_Nuclear_880_${i}_grouped.txt
done

for i in {1..1000};
do
Rscript groupRandomByChromHMM.R -f /project2/gilad/briana/threeprimeseq/data/random_QTLsnps/Total/chromHMM_overlap/randomres_overlapChromHMM_Total_118_${i}.txt -o /project2/gilad/briana/threeprimeseq/data/random_QTLsnps/Total/chromHMM_overlap_group/randomres_overlapChromHMM_Total_118_${i}_grouped.txt
done

Once I have the results I will paste the third column of each file together

cut -d$' ' -f 1,2 randomres_overlapChromHMM_Nuclear_880_1_grouped.txt > Nuc_chromOverlap.txt

for i in {1..1000};
do
paste -d" " Nuc_chromOverlap.txt <(cut -d" " -f 3 randomres_overlapChromHMM_Nuclear_880_${i}_grouped.txt) > tmp
mv tmp Nuc_chromOverlap.txt
done


cut -d$' ' -f 1,2 randomres_overlapChromHMM_Total_118_99_grouped.txt> Tot_chromOverlap.txt

for i in {1..1000};
do
paste -d" " Tot_chromOverlap.txt <(cut -d" " -f 3 randomres_overlapChromHMM_Total_118_${i}_grouped.txt) > tmp
mv tmp Tot_chromOverlap.txt
done

There will be NAs in this file. I will turn them into 0s when I bring it in R.

Pull files onto computer:

/project2/gilad/briana/threeprimeseq/data/random_QTLsnps/Nuclear/chromHMM_overlap_group/Nuc_chromOverlap.txt /project2/gilad/briana/threeprimeseq/data/random_QTLsnps/Total/chromHMM_overlap_group/Tot_chromOverlap.txt

regions=c('Txn_Elongation','Weak_Txn','Repressed','Heterochrom/lo','Repetitive/CNV1','Repetitive/CNV2','Active_Promoter','Weak_Promoter','Poised_Promoter','Strong_Enhancer1','Strong_Enhancer2','Weak_Enhancer1','Weak_Enhancer2','Insulator','Txn_Transition')


permutationResTotal=read.table("../data/ChromHmmOverlap/Tot_chromOverlap.txt", header=T, stringsAsFactors = F)
permutationResTotal[is.na(permutationResTotal)] <- 0
permutationResTotal= as_data_frame(permutationResTotal)
permutationResTotal_noName=permutationResTotal[,3:ncol(permutationResTotal)]
totRand_mean=rowMeans(permutationResTotal_noName)/1000

permutationResNuclear=read.table("../data/ChromHmmOverlap/Nuc_chromOverlap.txt",header = T,stringsAsFactors = F)
permutationResNuclear[is.na(permutationResNuclear)] <- 0
permutationResNuclear_noName=permutationResNuclear[,3:ncol(permutationResNuclear)]
nucRand_mean=rowMeans(permutationResNuclear_noName)/1000

allRand_mean_df= data.frame(cbind(regions,totRand_mean, nucRand_mean))

allRand_mean_df_melt=melt(allRand_mean_df, id.vars="regions")

Warning: attributes are not identical across measure variables; they will
be dropped

allRand_mean_df_melt$value= as.numeric(allRand_mean_df_melt$value)
ggplot(allRand_mean_df_melt, aes(y=value, x=regions, by=variable, fill=variable))+ geom_histogram(stat="identity", position="dodge") + theme(axis.text.x = element_text(angle = 90, hjust = 1))

Warning: Ignoring unknown parameters: binwidth, bins, pad

I want to look at specific distributions:

permutationResTotal_melt= melt(permutationResTotal, id.vars=c("number", "name"))

ggplot(permutationResTotal_melt, aes(x=value,fill=name)) + geom_histogram(bins=50) + facet_wrap(~number) + labs(x="N random Snps in category", title="Random permutation Total")

For nuclear:

permutationResNuclear_melt= melt(permutationResNuclear, id.vars=c("number", "name"))

ggplot(permutationResNuclear_melt, aes(x=value,fill=name)) + geom_histogram(bins=50) + facet_wrap(~number) + labs(x="N random Snps in category", title="Random permutation Nuclear")

Try log scale:

I want to add horizontal line where the actual QTL number is.

ggplot(permutationResTotal_melt, aes(x=value,fill=name)) + geom_histogram(bins=50) + facet_wrap(~number) + scale_y_log10()

Warning: Transformation introduced infinite values in continuous y-axis

Warning: Removed 448 rows containing missing values (geom_bar).

ggplot(permutationResNuclear_melt, aes(x=value,fill=name)) + geom_histogram(bins=50) + facet_wrap(~number) + scale_y_log10()

Warning: Transformation introduced infinite values in continuous y-axis

Warning: Removed 631 rows containing missing values (geom_bar).

Try removing 0s:

permutationResTotal_melt_no0= permutationResTotal_melt %>% filter(value>0)
ggplot(permutationResTotal_melt_no0, aes(x=value,fill=name)) + geom_histogram(bins=50) + facet_wrap(~number)

permutationResNuclear_melt_no0= permutationResNuclear_melt %>% filter(value>0)
ggplot(permutationResNuclear_melt_no0, aes(x=value,fill=name)) + geom_histogram(bins=50) + facet_wrap(~number)

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] grid      stats     graphics  grDevices utils     datasets  methods  
[8] base     

other attached packages:
 [1] bindrcpp_0.2.2      cowplot_0.9.3       ggpubr_0.1.8       
 [4] magrittr_1.5        data.table_1.11.8   VennDiagram_1.6.20 
 [7] futile.logger_1.4.3 forcats_0.3.0       stringr_1.3.1      
[10] dplyr_0.7.6         purrr_0.2.5         readr_1.1.1        
[13] tidyr_0.8.1         tibble_1.4.2        ggplot2_3.0.0      
[16] tidyverse_1.2.1     reshape2_1.4.3      workflowr_1.1.1    

loaded via a namespace (and not attached):
 [1] tidyselect_0.2.4     haven_1.1.2          lattice_0.20-35     
 [4] colorspace_1.3-2     htmltools_0.3.6      yaml_2.2.0          
 [7] rlang_0.2.2          R.oo_1.22.0          pillar_1.3.0        
[10] glue_1.3.0           withr_2.1.2          R.utils_2.7.0       
[13] RColorBrewer_1.1-2   lambda.r_1.2.3       modelr_0.1.2        
[16] readxl_1.1.0         bindr_0.1.1          plyr_1.8.4          
[19] munsell_0.5.0        gtable_0.2.0         cellranger_1.1.0    
[22] rvest_0.3.2          R.methodsS3_1.7.1    evaluate_0.11       
[25] labeling_0.3         knitr_1.20           broom_0.5.0         
[28] Rcpp_0.12.19         formatR_1.5          backports_1.1.2     
[31] scales_1.0.0         jsonlite_1.5         hms_0.4.2           
[34] digest_0.6.17        stringi_1.2.4        rprojroot_1.3-2     
[37] cli_1.0.1            tools_3.5.1          lazyeval_0.2.1      
[40] futile.options_1.0.1 crayon_1.3.4         whisker_0.3-2       
[43] pkgconfig_2.0.2      xml2_1.2.0           lubridate_1.7.4     
[46] assertthat_0.2.0     rmarkdown_1.10       httr_1.3.1          
[49] rstudioapi_0.8       R6_2.3.0             nlme_3.1-137        
[52] git2r_0.23.0         compiler_3.5.1