Problem with normalization by bcftools

Question

[Researcher]

I used the command below to normalize my vcf file by splitting multiple alternatives into multiple rows:

bcftools norm -m - <in.vcf> > out.vcf

but some of the variants disappeared after normalization!! Does anyone know why this happened?

[chrchang523]

Can you provide a specific example of disappearing variants?

[Researcher]

Here, the example:

Chromosome      2339162 .       C       T,A     89906.62        PASS    AC=25,1;                                                                             AF=0.114,4.566e-03;AN=219;BaseQRankSum=-3.610e-01;DP=17877;FS=1.706;MLEAC=25,1;M                                                                             LEAF=0.114,4.566e-03;MQ=60.00;MQRankSum=0.00;QD=34.36;ReadPosRankSum=0.412;SNPEF                                                                             F_AMINO_ACID_CHANGE=D87Y;SNPEFF_CODON_CHANGE=Gac/Tac;SNPEFF_EFFECT=NON_SYNONYMOU                                                                             S_CODING;SNPEFF_EXON_ID=1;SNPEFF_FUNCTIONAL_CLASS=MISSENSE;SNPEFF_GENE_BIOTYPE=p                                                                             rotein_coding;SNPEFF_GENE_NAME=gyrA;SNPEFF_IMPACT=MODERATE;SNPEFF_TRANSCRIPT_ID=                                                                             AAC75291;SOR=0.651      GT:AD:DP:GQ:PL  0:89,0,0:89:99:0,1800,1800      0:86,0,0                                                                             :86:99:0,1800,1800      0:78,0,0:78:99:0,1800,1800      0:98,0,0:98:99:0,1800,18                                                                             00      0:105,0,0:105:99:0,1800,1800    0:101,0,0:101:99:0,1800,1800    0:98,0,0                                                                          :98:99:0,1800,1800 

Btw, I didn't copy the genotype fields for all the samples.

[chrchang523]

That's not a complete example: you only stated the input, not the output and what was missing from it. Imagine yourself on the other side of this: what information would you need to reproduce the problem at all? If we can't do that, we can't really help.

[Researcher]

Because I can't post a long reply, I show only the variant that was part of the input, but it disappeared from the output. I hope this is helpful. Otherwise, which kind of information do you need?

[chrchang523]

Can you reproduce the problem by pasting just that line into a VCF file, and then copying over a bunch of header lines? If not, we have no hope; we don't even have access to the header lines which you haven't posted, or the bcftools version, or...