INFO/END< POS

[lindenb]

Hi,

FYI, when using /smoove-0.2.6 + singularity. With smoove genotype , some variants have POS>INFO/END.

command:

        singularity exec\
                --home \${PWD} \
                --bind ${bam.getParent()}:/bamdir \
                --bind ${vcf0.getParent()}:/mergeddir \
                --bind ${ref.getParent()}:/ref \
                --bind \${PWD}/TMP:/outdir \
                ${img} \
                smoove genotype -x  \
                        --vcf /mergeddir/${vcf0.name} \
                        --outdir /outdir \
                        --name ${sample} \
                        --fasta /ref/${ref.name} \
                        -p ${task.cpus} \
                        /bamdir/${bam.name}

log:

(..)
[smoove] 2021/06/01 10:16:57 Unpaired breakends found in file. These will not be present in output.
[smoove] 2021/06/01 10:17:21 [W::vcf_parse_info] INFO/END=60694778 is smaller than POS at chr2:60694786
[smoove] 2021/06/01 10:17:38 [W::tbx_parse1] VCF INFO/END=60694778 is smaller than POS at chr2:60694785
This tag will be ignored. Note: only one invalid END tag will be reported.

example:

chr2	60694786	23097	N	<DEL>	0	.	SVTYPE=DEL;SVLEN=8;END=60694778;STRANDS=+-:19;IMPRECISE;CIPOS=-342,65;CIEND=-31,228;CIPOS95=-299,65;CIEND95=-31,184;SU=19;PE=19;SR=0;SNAME=C001Q1M:404,C001KCJ:430;ALG=SUM	GT:GQ:SQ:GL:DP:RO:AO:QR:QA:RS:AS:ASC:RP:AP:AB	0/0:69:0:-0,-7,-23:23:23:0:23:0:23:0:0:0:0:0

the variant was already present in the 'merge' vcf:

chr2	60694786	23097	N	<DEL>	0	.	SVTYPE=DEL;SVLEN=8;END=60694778;STRA
NDS=+-:19;IMPRECISE;CIPOS=-342,65;CIEND=-31,228;CIPOS95=-299,65;CIEND95=-31,184;SU=19;PE=19;SR=0;PRP
OS=5.00000004629e-101,5.00000004629e-101,5.00000004629e-101,5.00000004629e-101,5.00000004629e-101,5.
00000004629e-101,5.00000004629e-101,5.00000004629e-101,5.00000004629e-101,5.00000004629e-101,5.00000
004629e-101,5.00000004629e-101,5.00000004629e-101,5.00000004629e-101,5.00000004629e-101,5.0000000462
9e-101,5.00000004629e-101,5.00000004629e-101,5.00000004629e-101,5.00000004629e-101,1.29049001195e-14
,1.48594501376e-13,1.71036001583e-12,1.4477650134e-11,1.22533001134e-10,1.0367600096e-09,8.772550081
21e-09,7.4207000687e-08,4.61517504273e-07,2.86851502656e-06,1.78328001651e-05,0.000110799001026,0.00
0201623001867,0.000366606003394,0.000666555006171,0.00121138001121,0.00220115502038,0.00294220002724
,0.00393070003639,0.00525145004862,0.00701510006494,0.00936785008673,0.0125010001157,0.0122675501136
,0.0120353001114,0.0118050501093,0.0115775001072,0.0113507001051,0.011124100103,0.0109008001009,0.01
06827000989,0.0104616000969,0.0102479500949,0.0100391500929,0.00982850009099,0.00961885008905,0.0094
1200008713,0.00920950008526,0.00900580008337,0.00880205008149,0.00860295007964,0.00841000007786,0.00
82196000761,0.00802755007432,0.00783575007254,0.00764660007079,0.00746275006909,0.00727920006739,0.0
0709835006572,0.00692030006407,0.00674400006243,0.00657360006086,0.00640310005928,0.00623425005772,0
.00607285005622,0.00591025005472,0.00575125005324,0.00559325005178,0.00543675005033,0.00528535004893
(...)

[brentp]

I'll get a fix for this in next release. Meanwhile, the solution is just to drop them. can do this with slivar with someting like:

slivar expr --pass-only  --info "variant.start <= variant.stop" -v $input -o $output

[fakedrtom]

Just following up here to see if this has been resolved in the latest release (0.2.8) as I am not seeing it detailed in the release notes...

[brentp]

no. not fixed. by next release, I mean, next release. ;) meanwhile, I would just add a step to remove these variants as they are low quality anyway.