Delete some INFO header lines and columns data

Question

[pablo.goiko]

Dear all

I have two vcf files I wish to concatenate, but they have different INFO tags. I would like to match the data in the two files by deleting the extra info in one of them.

I have tried "bcftools annotate", but I found it returns only the header, without the data. I have also read "Remove columns from the VCF file using vcftools" which is very promising...

awk -v OFS="\t" '!/##/ {$1=$2=$3=$4="";print}' test.vcf|sed 's/^\s\+//g'

Along the lines of this bash order, I wonder whether is it possible to discard specific INFO columns (lets say INFO/RPBZ, INFO/MQBZ and INFO/BQBZ) and the corresponding header lines. Does OFS admit more than one separator at once? If I could use OFS="\t" plus OFS=";", the above bash line could be easily modified to select the precise columns. In that case, how do I match the correct header lines?

Thanks in advance.

Pablo

[pablo.goiko]

Ooops.. Eventually I realized there is no need for a complex OFS and it can be accomplish with OFS=";". Sorry to bother your time.

    awk -v OFS=";" '!/##/ {$5=$6=$7=$8=$9=""}1' $file |sed 's/^\s\+//g' > $PRETWEAKBIFILE

bcftools annotate -x INFO/RPBZ,INFO/MQBZ,INFO/MQSBZ,INFO/BQBZ,INFO/SCBZ -O z -o $HEADFILE --threads 6 $file
bcftools reheader -h $HEADFILE -o TWEAKBIFILE --threads 6 $PRETWEAKBIFILE
bcftools index -t $TWEAKBIFILE

Pablo

[Pierre Lindenbaum]

this is useless, bcftools annotate already removes INFO from the header and from the variants.

[pablo.goiko]

Sorry Pierre... If that is ok with other users, fine. I double checked my "bcftools annotate infile" had the header and the variants but the outfile had only the header. Pablo

[Pierre Lindenbaum]

bcftools view rotavirus_rf.vcf.gz | grep VDB -m2
##INFO=<ID=VDB,Number=1,Type=Float,Description="Variant Distance Bias for filtering splice-site artefacts in RNA-seq data (bigger is better)",Version="3">
RF01    970 .   A   C   48.6696 .   DP=36;VDB=0.693968;SGB=10.3229;RPB=0.658863;MQB=1;MQSB=1;BQB=0.572843;MQ0F=0;ICB=0.425;HOB=0.32;AC=2;AN=10;DP4=19,7,3,5;MQ=60   GT:PL   0/0:0,9,47  0/0:0,18,73 0/0:0,18,73 0/0:0,33,116    1/1:95,24,0

remove 'INFO/VDB', the header INFO ID=VDB was removed. there is no line containing VDB but a reminder of the bcftools annotate command.

$ bcftools annotate -x 'INFO/VDB' rotavirus_rf.vcf.gz | grep VDB -m2
##bcftools_annotateCommand=annotate -x INFO/VDB rotavirus_rf.vcf.gz; Date=Sun Jan 30 16:35:35 2022

check the variant at 970 is still here but INFO/VDB was removed.

$ bcftools annotate -x 'INFO/VDB' rotavirus_rf.vcf.gz | awk '$2==970'
RF01    970 .   A   C   48.6696 .   DP=36;SGB=10.3229;RPB=0.658863;MQB=1;MQSB=1;BQB=0.572843;MQ0F=0;ICB=0.425;HOB=0.32;AC=2;AN=10;DP4=19,7,3,5;MQ=60    GT:PL   0/0:0,9,47  0/0:0,18,73 0/0:0,18,73 0/0:0,33,116    1/1:95,24,0

[pablo.goiko]

Thank you Pierre. I'll look forward my mistake(s).

Should I close the question? I have not found the right place to do it

[GenoMax]

There is no need to close the question. If @Pierre's answer below helped you can accept that (green checkmark) that provides closure to this thread. Do not delete the thread.

[pablo.goiko]

Thanks to everybody sharing knowledge!!

Answer 1

[Pierre Lindenbaum]

use bcftools annotate -x 'INFO/RPBZ,INFO/MQBZ,INFO/BQBZ' in.vcf

[pablo.goiko]

Thank you very much Pierre. In fact, that was my first option, with the -o option. But it returns me an outfile with just the header. Pablo