LNG + QMUL
- Project: Is there genetic overlap between MS and PD?
- Author(s): Cornelis B., Ben J., Frank G.
- Date Last Updated: December 2020
PD and MS are two brain disorders with both a potential immunesystem component, here we assess whether there is genetic overlap.
- Perform a genome-wide comparison between PD and MS sumstats...
- Perform a localized comparison between PD and MS GWAS regions...
- ...
TBD
TBD
Using LDSCORE to assess genome-wide correlation
cd /data/CARD/projects/MS_PD/
MS: using stats => discovery_metav3.0.meta.gz
stats from here => https://pubmed.ncbi.nlm.nih.gov/31604244/ data obtained from https://imsgc.net/
PD: using stats => /data/CARD/PD/summary_stats/resultsForSmr_filtered.tab.gz
stats from here => https://pubmed.ncbi.nlm.nih.gov/31701892/ using data plus 23andMe and UK Biobank
module load ldsc/1.0.0-101-g89c13a7
# get reference data
wget https://data.broadinstitute.org/alkesgroup/LDSCORE/eur_w_ld_chr.tar.bz2
tar -jxvf eur_w_ld_chr.tar.bz2
# PD
munge_sumstats.py --out PD_all_LDSC --sumstats /data/CARD/PD/summary_stats/resultsForSmr_filtered.tab.gz \
--merge-alleles eur_w_ld_chr/w_hm3.snplist \
--snp SNP --a1 A1 --a2 A2 --p p --frq freq
# MS
## remove N column
zless discovery_metav3.0.meta.gz | cut -d " " -f1,2,3,4,5,7,8 | grep rs > temp.txt
zless discovery_metav3.0.meta.gz | cut -d " " -f1,2,3,4,5,7,8 | head -1 > header.txt
cat header.txt temp.txt > input_MS.txt
munge_sumstats.py --out MS_v3_LDSC --sumstats input_MS.txt \
--merge-alleles eur_w_ld_chr/w_hm3.snplist \
--N 115000 --snp SNP --a1 A1 --a2 A2 --p P
# do final comparison
ldsc.py --rg PD_all_LDSC.sumstats.gz,MS_v3_LDSC.sumstats.gz --out PD_MS --ref-ld-chr eur_w_ld_chr/ --w-ld-chr eur_w_ld_chr/
# summary
Summary of Genetic Correlation Results
p1 p2 rg se z p h2_obs h2_obs_se h2_int h2_int_se gcov_int gcov_int_se
PD_all_LDSC.sumstats.gz MS_v3_LDSC.sumstats.gz -0.0027 0.029 -0.0922 0.9266 0.1016 0.01 1.0215 0.0108 0.0474 0.005
# no real correlation to be seen here...
Taking sumstats from here https://pubmed.ncbi.nlm.nih.gov/31604244/
Using Sup Table 7 (only autosomals) => 197 signals
Using PD GWAS hits from GWAS browser => Nalls et al 2019 https://pdgenetics.shinyapps.io/GWASBrowser/
cd /data/CARD/projects/MS_PD/
module load bedtools
cut -f 1-3 MS_GWAS.bed > MS_short.bed #plus remove header
cut -f 1-3 PD_GWAS.bed > PD_short.bed #plus remove header
bedtools window -w 500000 -a MS_short.bed -b PD_short.bed
MS-SIDE | PD SIDE | Distance (MS-PD)
1 154983036 154983036 > ZBTB7B area | 1 154898185 154898185 > GBA area | 84851
1 154983036 154983036 > ZBTB7B area | 1 155135036 155135036 > GBA area | -152000
1 154983036 154983036 > ZBTB7B area | 1 155205634 155205634 > GBA area | -222598
3 18798848 18798848 > SATB1 area | 3 18361759 18361759 => SATB1 area | 437089
3 121765368 121765368 > CD86 | 3 122196892 122196892 => KPNA1 area | -431524
3 121783015 121783015 > CD86 | 3 122196892 122196892 => KPNA1 area | -413877
5 133891282 133891282 > JADE2 | 5 134199105 134199105 => C5orf24 area | -307823
12 123604053 123604053 > PITPNM2 | 12 123326598 123326598 => HIP1R area | 277455
14 88523488 88523488 > LINC01146 | 14 88464264 88464264 => GALC area | 59224
17 40529835 40529835 > STAT3 | 17 40741013 40741013 => RETREG3 area | -211178
17 43407670 43407670 > MAP3K14 | 17 43744203 43744203 => MAPT area | -336533
17 43407670 43407670 > MAP3K14 | 17 43798308 43798308 => MAPT area | -390638
# LD link comparison MS + PD => R2 = XX D' = XX
--chr 1 region ZBTB7B:
rs112344141 + rs114138760 => R2 = 0.0006 D' = 0.0454
rs112344141 + rs35749011 => R2 = 0.0009 D' = 1.0 # warning low freq variant
rs112344141 + rs76763715 => R2 = 0.0001 D' = 1.0 # warning low freq variant
--chr 3 region SATB1 :
rs9863496 + rs73038319 => R2 = 0.0154 D' = 0.3911
--chr 3 region CD86 :
rs200866143 + rs55961674 => R2 = XX D' = XX # rs200866143 not in 1K genomes
rs75937181 + rs55961674 => R2 = 0.002 D' = 0.0634
--chr 5 region JADE2 :
rs2084007 + rs11950533 => R2 = 0.1159 D' = 0.8675
--chr 12 region PITPNM2 :
rs7975763 + rs10847864 => R2 = 0.0294 D' = 0.4562
--chr 14 region LINC01146 :
rs116899835 + rs979812 => R2 = 0.0669 D' = 0.9654
--chr 17 region STAT3 :
rs1026916 + rs12951632 => R2 = 0.0849 D' = 0.3681
--chr 17 region MAP3K14 :
rs7222450 + rs62053943 => R2 = 0.0011 D' = 0.0777
rs7222450 + rs117615688 => R2 = 0.0142 D' = 0.4597
# nothing really convincing to be honest....
Also a correlation between P-values using locus-compare => http://www.locuscompare.com/
cd /data/CARD/projects/MS_PD/
MS: using stats => discovery_metav3.0.meta.gz
stats from here => https://pubmed.ncbi.nlm.nih.gov/31604244/ data obtained from https://imsgc.net/
PD: using stats => /data/CARD/PD/summary_stats/resultsForSmr_filtered.tab.gz
stats from here => https://pubmed.ncbi.nlm.nih.gov/31701892/ using data plus 23andMe and UK Biobank
need format:
rsid pval
rs6933892 0.243779
rs6934084 0.239798
rs210910 0.193318
# GALC area only...
14 86000000 90000000 => GALC area
awk '{ if($2 >= 86000000) { print }}' pos_cut.txt | head
## MS GALC only...
zless discovery_metav3.0.meta.gz | awk '{ if($1 == 14) { print }}' | awk '{ if($2 >= 86000000) { print }}' | awk '{ if($2 <= 90000000) { print }}' \
| cut -d " " -f 3,7 | grep rs | grep -v NA | sed -e 's/ /\t/g' > temp.txt
echo "rsid" > part1.txt
echo "pval" > part2.txt
paste part1.txt part2.txt > header2.txt
cat header2.txt temp.txt > MS_input_locuscompare.txt
## MS
zless discovery_metav3.0.meta.gz | cut -d " " -f 3,7 | grep rs | grep -v NA | sed -e 's/ /\t/g' > temp.txt
echo "rsid" > part1.txt
echo "pval" > part2.txt
paste part1.txt part2.txt > header2.txt
cat header2.txt temp.txt > MS_input_locuscompare.txt
## PD
module load R
R
require(data.table)
PD <- fread("/data/CARD/PD/summary_stats/resultsForSmr_filtered.tab.gz",header=T)
MS <- fread("MS_input_locuscompare.txt",header=T)
MM <- merge(PD,MS,by.x="SNP",by.y="rsid")
write.table(MM, file="PD_to_Edit_input.txt", quote=FALSE,row.names=F,sep="\t")
# download and upload....
