/
snpinSS.Rmd
255 lines (146 loc) · 13 KB
/
snpinSS.Rmd
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
---
title: "SNPs in signal sites"
author: "Briana Mittleman"
date: "6/24/2019"
output: html_document
---
```{r setup, include=FALSE}
knitr::opts_chunk$set(echo = TRUE)
```
In this analysis I want to ask if snps in a signal site are more likely to be apaQTLs than other snps close to the PAS. In order to do this i need to subset to the pas that have signal site (identified [here](signalsiteanalysis.html)) I will then identyify the region 50 bp upstream of the PAS and ask if there are snps in this region using the vcf files for the snps i tested.
```{r}
library(workflowr)
library(tidyverse)
library(ggpubr)
```
```{bash,eval=F}
mkdir ../data/SNPinSS
```
I want a bed file with 50bp upstream of these PAS.
```{r}
PASwSS=read.table("../data/PAS/PASwSignalSite.txt", header = T,stringsAsFactors = F)
PAS=read.table("../data/PAS/APAPAS_GeneLocAnno.5perc_withCHR.bed", stringsAsFactors = F, header = F, col.names = c("chr", "start", "end", "PASid","score", "strand")) %>% separate(PASid, into=c("pasNum", "geneiD"), sep=":") %>% mutate(PAS=paste("peak", pasNum, sep=""),PASname=paste(PAS, geneiD, sep="_"))
PASwSSregion=PASwSS %>% inner_join(PAS, by="PAS") %>% mutate(newEnd=ifelse(strand=="+", end+50, end),newStart=ifelse(strand=="+", start, start-50)) %>% select(chr, newStart,newEnd, PASname, score, strand)
write.table(PASwSSregion,"../data/SNPinSS/FiftyupstreamPASwSS.bed", col.names = F, row.names = F, quote = F, sep="\t")
```
```{bash,eval=F}
sed 's/^chr//' ../data/SNPinSS/FiftyupstreamPASwSS.bed > ../data/SNPinSS/FiftyupstreamPASwSS.nochr.bed
sort -k1,1 -k2,2n ../data/SNPinSS/FiftyupstreamPASwSS.nochr.bed > ../data/SNPinSS/FiftyupstreamPASwSS.nochr.sort.bed
sbatch subsetVCF_upstreamPAS.sh
cat ../data/SNPinSS/SNPSinFiftyupstreamPAS_chr* >../data/SNPinSS/SNPSinFiftyupstreamPAS_Allchr.recode.vcf
```
I want to further subset to those in a signal site.
```{r}
SSregions=PASwSS %>% inner_join(PAS, by="PAS") %>% mutate(absdist=abs(UpstreamDist),newEnd= ifelse(strand=="+", end-absdist, end+absdist), newStart=ifelse(strand=="+", end- (absdist+6), end + (absdist-6)), length=newEnd-newStart) %>% select(chr, newStart,newEnd, PASname, score, strand)
write.table(SSregions,"../data/SNPinSS/SignalSiteRegions.bed", col.names = F, row.names = F, quote = F, sep="\t")
```
```{bash,eval=F}
sed 's/^chr//' ../data/SNPinSS/SignalSiteRegions.bed > ../data/SNPinSS/SignalSiteRegions.nochr.bed
sort -k1,1 -k2,2n ../data/SNPinSS/SignalSiteRegions.nochr.bed > ../data/SNPinSS/SignalSiteRegions.nochr.sort.bed
sbatch subsetVCF_SS.sh
cat ../data/SNPinSS/SSregions_chr* > ../data/SNPinSS/SSregions_Allchr.recode.vcf
```
I will also need a different region to comare to. I can just shift these regions upstream by 7
```{r}
SS_diffregion=SSregions %>% mutate(randStart=ifelse(strand=="+", newStart-7, newEnd), randend=ifelse(strand=="+", newStart, newEnd+7), length=randend-randStart) %>% select(chr, randStart,randend, PASname, score, strand)
write.table(SS_diffregion,"../data/SNPinSS/OtherSSRegions.bed", col.names = F, row.names = F, quote = F, sep="\t")
```
```{bash,eval=F}
sed 's/^chr//' ../data/SNPinSS/OtherSSRegions.bed > ../data/SNPinSS/OtherSSRegions.nochr.bed
sort -k1,1 -k2,2n ../data/SNPinSS/OtherSSRegions.nochr.bed > ../data/SNPinSS/OtherSSRegions.nochr.sort.bed
sbatch subsetvcf_otherreg.sh
cat ../data/SNPinSS/Otherregions_chr* > ../data/SNPinSS/Otherregions_Allchr.recode.vcf
```
Alternative option, permute the distances:
```{r}
permdist=sample(PASwSS$UpstreamDist, length(PASwSS$UpstreamDist), replace = F)
SSregions_perm=as.data.frame(cbind(PASwSS, permdist))%>% inner_join(PAS, by="PAS") %>% mutate(absdist=abs(permdist),newEnd= ifelse(strand=="+", end-absdist, end+absdist), newStart=ifelse(strand=="+", end- (absdist+6), end + (absdist-6)), length=newEnd-newStart)%>% select(chr, newStart,newEnd, PASname, score, strand)
write.table(SSregions_perm,"../data/SNPinSS/SSRegions_permuted.bed", col.names = F, row.names = F, quote = F, sep="\t")
```
```{bash,eval=F}
sed 's/^chr//' ../data/SNPinSS/SSRegions_permuted.bed > ../data/SNPinSS/SSRegions_permuted.nochr.bed
sort -k1,1 -k2,2n ../data/SNPinSS/SSRegions_permuted.nochr.bed > ../data/SNPinSS/SSRegions_permuted.nochr.sort.bed
sbatch subsetvcf_permSS.sh
cat ../data/SNPinSS/SSRegionsPerm_chr* > ../data/SNPinSS/SSRegionsPerm_Allchr.recode.vcf
#remove # in first line
```
Pull in QTL snps:
```{r}
totQTLs=read.table("../data/apaQTLs/Total_apaQTLs4pc_5fdr.txt",header = T, stringsAsFactors = F) %>% select(sid) %>% unique()
write.table(totQTLs,"../data/apaQTLs/TotalQTLSNPsRSID.txt", col.names = F, row.names = F, quote = F)
nucQTLs=read.table("../data/apaQTLs/Nuclear_apaQTLs4pc_5fdr.txt",header = T, stringsAsFactors = F) %>% select(sid) %>% unique()
write.table(nucQTLs,"../data/apaQTLs/NuclearQTLSNPsRSID.txt", col.names = F, row.names = F, quote = F)
```
Signal site results:
```{r}
SS_snps=read.table("../data/SNPinSS/SSregions_Allchr.recode.vcf",header = T, stringsAsFactors = F) %>% select(ID) %>% mutate(totQTL=ifelse(ID %in% totQTLs$sid, "Yes", "No"), nucQTL=ifelse(ID %in% nucQTLs$sid, "Yes", "No"))
permutedSS_snps=read.table("../data/SNPinSS/SSRegionsPerm_Allchr.recode.vcf",header = T, stringsAsFactors = F) %>% select(ID) %>% mutate(totQTL=ifelse(ID %in% totQTLs$sid, "Yes", "No"), nucQTL=ifelse(ID %in% nucQTLs$sid, "Yes", "No"))
otherReg_snp=read.table("../data/SNPinSS/Otherregions_Allchr.recode.vcf",header = T, stringsAsFactors = F) %>% select(ID) %>% mutate(totQTL=ifelse(ID %in% totQTLs$sid, "Yes", "No"), nucQTL=ifelse(ID %in% nucQTLs$sid, "Yes", "No"))
fiftybp_snp=read.table("../data/SNPinSS/SNPSinFiftyupstreamPAS_Allchr.recode.vcf",header = T, stringsAsFactors = F) %>% select(ID) %>% mutate(totQTL=ifelse(ID %in% totQTLs$sid, "Yes", "No"), nucQTL=ifelse(ID %in% nucQTLs$sid, "Yes", "No"))
```
There are only 2 qtl snps in these signal sites. This is not enough to draw anything from this.
Try with pvalues. Are the snps in pvals more likely to be significant than those not.
I need to figure out which peak is associated with each snp.
I can make a bedfile from the SS snps in python and overlap this with the Signal site regions.
```{bash,eval=F}
python vcf2bed.py ../data/SNPinSS/SSregions_Allchr.recode.vcf ../data/SNPinSS/SSregions_Allchr.bed
sort -k1,1 -k2,2n ../data/SNPinSS/SSregions_Allchr.bed > ../data/SNPinSS/SSregions_Allchr.sort.bed
python vcf2bed.py ../data/SNPinSS/SSRegionsPerm_Allchr.recode.vcf ../data/SNPinSS/SSRegionsPerm_Allchr.bed
sort -k1,1 -k2,2n ../data/SNPinSS/SSRegionsPerm_Allchr.bed> ../data/SNPinSS/SSRegionsPerm_Allchr.sort.bed
python vcf2bed.py ../data/SNPinSS/SNPSinFiftyupstreamPAS_Allchr.recode.vcf ../data/SNPinSS/SNPSinFiftyupstreamPAS_Allchr.bed
sort -k1,1 -k2,2n ../data/SNPinSS/SNPSinFiftyupstreamPAS_Allchr.bed > ../data/SNPinSS/SNPSinFiftyupstreamPAS_Allchr.sort.bed
python vcf2bed.py ../data/SNPinSS/Otherregions_Allchr.recode.vcf ../data/SNPinSS/Otherregions_Allchr.bed
sort -k1,1 -k2,2n ../data/SNPinSS/Otherregions_Allchr.bed > ../data/SNPinSS/Otherregions_Allchr.sort.bed
sort -k1,1 -k2,2n ../data/SNPinSS/SignalSiteRegions.bed > ../data/SNPinSS/SignalSiteRegions.sort.bed
sort -k1,1 -k2,2n ../data/SNPinSS/SSRegions_permuted.bed >../data/SNPinSS/SSRegions_permuted.sort.bed
sort -k1,1 -k2,2n ../data/SNPinSS/FiftyupstreamPASwSS.bed > ../data/SNPinSS/FiftyupstreamPASwSS.sort.bed
sort -k1,1 -k2,2n ../data/SNPinSS/OtherSSRegions.bed > ../data/SNPinSS/OtherSSRegions.sort.bed
```
intersect with bedtools to map the snps to the regions. Then I will be able to select the snp PAS associations.
```{bash,eval=F}
sbatch mapSSsnps2PAS.sh
```
Results to get the associations:
```{r}
SSsnpswithPAS=read.table("../data/SNPinSS/SNPinSS2PAS.txt",col.names = c("chr","start", "end", "PASname", "score", "strand", "SNP")) %>% filter(SNP!=".") %>% separate(PASname, into=c("PAS", "gene", "loc"),sep="_") %>% select(PAS, SNP)
write.table(SSsnpswithPAS, "../data/SNPinSS/SS_PASandSNPs.txt", row.names = F, col.names = F, quote = F, sep="\t")
SSsnpswithPERMPAS=read.table("../data/SNPinSS/SNPinPermSS2PAS.txt",col.names = c("chr","start", "end", "PASname", "score", "strand", "SNP")) %>% filter(SNP!=".") %>% separate(PASname, into=c("PAS", "gene", "loc"),sep="_") %>% select(PAS, SNP)
write.table(SSsnpswithPERMPAS, "../data/SNPinSS/PermSS_PASandSNPs.txt", row.names = F, col.names = F, quote = F, sep="\t")
SNPregion=read.table("../data/SNPinSS/SNPSinFiftyupstream2PAS.txt",col.names = c("chr","start", "end", "PASname", "score", "strand", "SNP")) %>% filter(SNP!=".") %>% separate(PASname, into=c("PAS", "gene", "loc"),sep="_") %>% select(PAS, SNP)
write.table(SNPregion, "../data/SNPinSS/PASregion_PASandSNPs.txt", row.names = F, col.names = F, quote = F, sep="\t")
Otherregion=read.table("../data/SNPinSS/Otherregions2PAS.txt",col.names = c("chr","start", "end", "PASname", "score", "strand", "SNP")) %>% filter(SNP!=".") %>% separate(PASname, into=c("PAS", "gene", "loc"),sep="_") %>% select(PAS, SNP)
write.table(Otherregion, "../data/SNPinSS/Otherregions_PASandSNPs.txt", row.names = F, col.names = F, quote = F, sep="\t")
```
Process the snp region with python
```{bash,eval=F}
python fixPASregionSNPs.py
```
Run this with total and nuclear
```{bash,eval=F}
python NomResfromPASSNP.py ../data/SNPinSS/SS_PASandSNPs.txt ../data/apaQTLNominal_4pc/APApeak_Phenotype_GeneLocAnno.Nuclear.5perc.fc.gz.qqnorm_AllChrom.txt ../data/apaQTLNominal_4pc/SS_Nuclear_nomRes.txt
python NomResfromPASSNP.py ../data/SNPinSS/SS_PASandSNPs.txt ../data/apaQTLNominal_4pc/APApeak_Phenotype_GeneLocAnno.Total.5perc.fc.gz.qqnorm_AllChrom.txt ../data/apaQTLNominal_4pc/SS_Total_nomRes.txt
python NomResfromPASSNP.py ../data/SNPinSS/PermSS_PASandSNPs.txt ../data/apaQTLNominal_4pc/APApeak_Phenotype_GeneLocAnno.Nuclear.5perc.fc.gz.qqnorm_AllChrom.txt ../data/apaQTLNominal_4pc/PermSS_Nuclear_nomRes.txt
python NomResfromPASSNP.py ../data/SNPinSS/PermSS_PASandSNPs.txt ../data/apaQTLNominal_4pc/APApeak_Phenotype_GeneLocAnno.Total.5perc.fc.gz.qqnorm_AllChrom.txt ../data/apaQTLNominal_4pc/PermSS_Total_nomRes.txt
python NomResfromPASSNP.py ../data/SNPinSS/PASregion_PASandSNPs.FIXED.txt ../data/apaQTLNominal_4pc/APApeak_Phenotype_GeneLocAnno.Nuclear.5perc.fc.gz.qqnorm_AllChrom.txt ../data/apaQTLNominal_4pc/RegionSS_Nuclear_nomRes.txt
python NomResfromPASSNP.py ../data/SNPinSS/PASregion_PASandSNPs.FIXED.txt ../data/apaQTLNominal_4pc/APApeak_Phenotype_GeneLocAnno.Total.5perc.fc.gz.qqnorm_AllChrom.txt ../data/apaQTLNominal_4pc/RegionSS_Total_nomRes.txt
python NomResfromPASSNP.py ../data/SNPinSS/Otherregions_PASandSNPs.FIXED.txt ../data/apaQTLNominal_4pc/APApeak_Phenotype_GeneLocAnno.Nuclear.5perc.fc.gz.qqnorm_AllChrom.txt ../data/apaQTLNominal_4pc/OtherSS_Nuclear_nomRes.txt
python NomResfromPASSNP.py ../data/SNPinSS/Otherregions_PASandSNPs.FIXED.txt ../data/apaQTLNominal_4pc/APApeak_Phenotype_GeneLocAnno.Total.5perc.fc.gz.qqnorm_AllChrom.txt ../data/apaQTLNominal_4pc/OtherSS_Total_nomRes.txt
```
```{r}
Nuclear_SS=read.table('../data/apaQTLNominal_4pc/SS_Nuclear_nomRes.txt', header = F, col.names=c("peakID", "snp", "dist", "pval", "slope"), stringsAsFactors = F) %>% select(pval,slope) %>% mutate(fraction="Nuclear", set="SS")
Nuclear_Perm=read.table('../data/apaQTLNominal_4pc/PermSS_Nuclear_nomRes.txt',header = F, col.names=c("peakID", "snp", "dist", "pval", "slope"), stringsAsFactors = F) %>% select(pval,slope) %>% mutate(fraction="Nuclear", set="Permuted")
Total_SS=read.table('../data/apaQTLNominal_4pc/SS_Total_nomRes.txt', header = F, col.names=c("peakID", "snp", "dist", "pval", "slope"), stringsAsFactors = F) %>% select(pval,slope) %>% mutate(fraction="Total", set="SS")
Total_Perm=read.table('../data/apaQTLNominal_4pc/PermSS_Total_nomRes.txt',header = F, col.names=c("peakID", "snp", "dist", "pval", "slope"), stringsAsFactors = F) %>% select(pval,slope) %>% mutate(fraction="Total", set="Permuted")
Nuclear_Region=read.table('../data/apaQTLNominal_4pc/RegionSS_Nuclear_nomRes.txt',header = F, col.names=c("peakID", "snp", "dist", "pval", "slope"), stringsAsFactors = F) %>% select(pval,slope) %>% mutate(fraction="Nuclear", set="Region")
Total_Region=read.table('../data/apaQTLNominal_4pc/RegionSS_Total_nomRes.txt',header = F, col.names=c("peakID", "snp", "dist", "pval", "slope"), stringsAsFactors = F) %>% select(pval,slope) %>% mutate(fraction="Total", set="Region")
Nuclear_other=read.table('../data/apaQTLNominal_4pc/OtherSS_Nuclear_nomRes.txt',header = F, col.names=c("peakID", "snp", "dist", "pval", "slope"), stringsAsFactors = F) %>% select(pval,slope) %>% mutate(fraction="Nuclear", set="Upstream")
Total_other=read.table('../data/apaQTLNominal_4pc/OtherSS_Total_nomRes.txt',header = F, col.names=c("peakID", "snp", "dist", "pval", "slope"), stringsAsFactors = F) %>% select(pval,slope) %>% mutate(fraction="Total", set="Upstream")
all_SS_pval=bind_rows(Nuclear_SS,Nuclear_Perm,Total_SS,Total_Perm,Nuclear_Region,Total_Region,Nuclear_other,Total_other)
```
```{r}
ggplot(all_SS_pval, aes(x=fraction, fill=set, y=pval)) + geom_boxplot()
```
```{r}
ggplot(all_SS_pval, aes(x=fraction, fill=set, y=abs(slope))) + geom_boxplot() + labs(y="absolute value effect size")
```
Does not look like there is an effect here.