Minimac3 to plink using GCTA

Question

[njbernstein]

Hi all,

I'm trying to convert minimal dosages into beds using gcta. Here's my code:

gcta64 --dosage-mach chr22.imputed.dose chr22.imputed.info --make-bed --out chr22.fdgsdf

My error is:

Reading dosage data failed. Please check the format of the map file.

Here's what my info file looks like:

SNP    REF    ALT    Major    Minor    MAF    AvgCall    Rsq    Genotyped    LooRsq    EmpR    EmpRsq    Dose1    Dose2
22:16050115    G    A    G    A    0.00639    0.99361    0.00000    -    -    -    -    -    -
22:16050213    C    T    C    T    0.00759    0.99241    0.00000    -    -    -    -    -    -
22:16050568    C    A    C    A    0.00040    0.99960    0.00005    -    -    -    -    -    -
22:16050607    G    A    G    A    0.00100    0.99900    0.00002    -    -    -    -    -    -
22:16050627    G    T    G    T    0.00040    0.99960    0.00005    -    -    -    -    -    -
22:16050654    A    <CN0>    A    <CN0>    0.00180    0.99820    0.00000    -    -    -    -    -    -
22:16050654    A    <CN2>    A    <CN2>    0.01737    0.98263    0.00000    -    -    -    -    -    -

Any ideas?

Best

Answer 1

[njbernstein]

Answer:

for chr in `seq 1 22`;
do
awk -v OFS='\t'  '!($4="") , !($5="")' chr$chr.imputed.info > chr$chr.imputed.info.reworked
sed -i 's/\t\t//' chr$chr.imputed.info.reworked 
awk '{t=$5; $5=$4; $4=t; print}' chr$chr.imputed.info.reworked  > t && mv t chr$chr.imputed.info.reworked
awk '{$6=.999 FS $6;}1' OFS='\t'  chr$chr.imputed.info.reworked > t && mv t chr$chr.imputed.info.reworked
awk '$1=$1"_"$2"_"$3"_"++i' chr$chr.imputed.info.reworked > t && mv t chr$chr.imputed.info.reworked
sed -i '1 s/^.*$/SNP\tAl1\tAl2\tFreq1\tMAF\tAvgCall\tRsq\tGenotyped\tLooRsq\tEmpR\tEmpRsq\tDose1\tDose2/g' chr$chr.imputed.info.reworked

gzip chr$chr.imputed.info.reworked
/labseq/tools/gcta_1.24.2/gcta64  --dosage-mach-gz chr$chr.imputed.dose.gz chr$chr.imputed.info.reworked.gz  --make-bed --out chr$chr.imputed
done

[xliu4255]

Hi thanks for the answer. I'm using GCTA to convert Minimac3 output files as well. Just want to ask have you using filter option to filter out low rsqr genotypes? Using the --dosage-mach command, and adding the filter flag, I found that it actually filter out the wrong column--- the AvgCall rather than R2. Have you encountered a similar problem? Thanks