Zero position in VCF

[drtconway]

Hi Noodles,

I think this might be a bug in Noodles. At first I thought it was a bug in the upstream SV caller, but on reflection, I think they might be doing the right thing; well more or less.

I have a SV that begins:

chr4_GL000008v2_random  0   cuteSV.INS.8444 G   GGTACTTTTTGTTTCTCAG...

It's a corner case that the VCF specification doesn't handle gracefully.

It's seems clear that the intent is that the ALT sequence should be inserted before the beginning of the contig.

Now there is obviously the problem of the context base in the REF column - putting the first base G is kind of misleading - I posit that using N would be preferable.

However my main concern is with the 0 position. The VCF parser gives a None for rec.variant_start(), which given that positions are 1-based, is kind of reasonable, except for this specific edge case, where 0 would actually be more accurate.

The result of rec.variant_start() is easy to deal with - I can map None onto 0 and get on with things.

Where things run into trouble is if I call rec.variant_end() which produces an error:

invalid INFO END position

If rec.variant_start() returns None for this case, would it be more consistent for rec.variant_end() to do the same here?

It's a bit of an annoying corner case!

Cheers,
Tom.

[drtconway]

Also, we need to be able to construct records for these cases.

[zaeleus]

Are you able to provide a complete example record or at least the INFO END value? I suspect the error is from an INFO END value <= 0, which would indeed be invalid/nonsensical.

[drtconway]

Here's a complete record:

chr4_GL000008v2_random	0	cuteSV.INS.8444	G	GGTACTTTTTGTTTCTCAGTGACTTGTGACTATTAACAAAAAGAGAGAGAGAGAGTAAAAAAGGAAAAAAACCTCTACTTCAGAGTGAGACAACGGGTCCGACCTCACGTTACCGTTGTAGAGACAAGTGACGTTGGAAGTGTAGGACCCAAGTCCGCTAAAAGGACGGAGTCGGAGGGTTCATCGACCCTAAGGTCCACGTAGGGCGGTACGGACCTATTCAAAAACATAAAAGTCATCTCTGCCCCAAAGTAGTACAACCGGTCCGACCAGAGTTTGAGGACTGGAGTTTACTAGACGGGCGGAATCGGAGGGTTTAACGACCACTACTAAAAGTAAATTACCACAAGATGTTTCACTCAAATCCGTTACTTTCGCGTGTACAGTAGATATCAAATGGTCTTTCATCATGGCCAGTGGTCGTAACTTCGTATGAGGATTGTGACCTTTTAAAGAAGTCGTCGGTCTTAATTAGAGTCTAAAGCTCTGACCTCGTTTCTGATTAAACTATACACAAGTAAGTCTAATCATTTAGAGGTACCACAAACTTGACATATTATAGTTTAGATTTTATTCAGTTTTCTCAACTTGTTATAGAGTAGAGACTCTTTTTTTACGTAATAAAAGTTGTCTAGATTCTGACTTAAATTTTAGGTTGACGGATAAGGTCACAAACAAAGAGATCTTATTCCGGTTCGGGATGCTTAAGACCTATCACAGGAGGTTCAACATAAATTCACTATTAACTATTTAGATACCAAATCTATGAATACCTTAAGGTATATTGAGATCGAGGTCGGAAACACTCATTTTTAACCGTCACTAGTACCGGAATTACTCCCGTACACCAATTAACCTTAAAAGTCATGGTTTTCAGCAGAAGAATGATCAAATAGTCCCATGGGATATTATCATAGGTATTGAGTAACGGGTGAACAGACAAAACTAAAGAAAAAATTTTTCTGTTAAAATTTGAATCTCCTAAACATGGAGTTCAGTGTATACACCGTGTAAATATTTTATTAGTACGGAAACGTATAAAAAGGAAACTATTGAATCGATATATAGGTTCGGTAAATACAAAAAATTAACAGTGTTTATAACGTGACTTTATCTCGTGATCAACGTTTTTACGGACCACATTGTAAAGAAGAACAAGAACTTTCAATAAAATTATGGGGTACAAATTCTATAAGGATCTGTTGATATACATTGATATTAATAAAGACGATCCATCTTTGTAGAAACACCAGTTGTAAACACTTCACGGAACATTTATTTTTATACGAACTCAACTTTATCATAAAACGTTTATATCGGGATTTATCTCGATGACTAAAATGGTTCTTAATTTCTGAGGTTTAAGACAGAAACAGAAAGTTTAATCCAAATATCCTTGTAGAGTAATGTTACTCGGTTTTCGGATAGACTAGTAATTACTCTATTAAAACTGTGAAGGTTTACTGTGTCAATTAAACATCATCCAAAAACTTCACGTTTAATGTACAAAGACAAATTTATTACTGTCATTTTATACCTGTGATAGATAAGTGTGAGATATATGTGTTGATGTTGAACAGAGATCTTCTTTTCAAATTAAAAAGTCCAGAGTAGGTGTTTTACTTCACAATTATTCGACATAGAAATAAGAGTAAAAATTCATTAAAGAAGAATTGGTAACATATTGTGATGGTGTCAAACATACCTCTATTTTATTGAGTACCACTCTTTAGGTCAAGAAATTTTTAGTAAAATGGAAGTTAGTGAAAATCCTATAGATACCAACATCTCATTCCTACTTTTAAATATTTGTAATCAATTTAATACGTTGACAAAAACTTCTACAAAGTTTTAATGTAAAAGGAATAAAATGTTGATCAAAAGTACATAAGTGTTTACTCTGTGTACTCTTAATAATTACGTTTTCGTGTCGCATAGTTGTTTCAAATATTACATTTATCAAATCAATCCTTTCTTACATTACTCACTTAATAGTTTCGATAAAATAACCAAGGATCTAAATGGTATCTTCTTTTATTTCGTCGGAATATTTTTAAAGAGTTGTGACTCAGGTAAATTCCAGATAAAACATACGGCATTCAAGATTTTTTACAATAACGGCCATTTTTATCCTACGGATAGATTTTAACAGAAAAATTTCCAAGAAAGAATAAGAAGAAGGAAAGGATACACACGTGTCTTATGACAGTAGTAAAACTAACCTTGTGACAGAGAGAAAGAGTCTCCCATATGTAAACATCTTCCCTCTTTTATAAACAATTAAAACGAAGGATATTCTATTCCAGAGTTACCAACTTCCAGTCAATTGTTTGTAAAAAAAAACAGAAGTATGATAATAGAACCTTAATGGTTGACCGAAGATAAATCTAAAACTTAATACTTTCTTTTCTTGATCGTAAACAATCTTGAATGACACAAGGTCCTTGAAATAACTATGGTAAAGAAGAGTAATGGTGTAAACAAGAAAGTTAAGCAGAATAAAAAAAGGTAGAATTATTCCACCCTGGTCTCCAAAACATTATAGGTGATTGGGTGACACCTAAAGTACCGAACGTTCTCCTTCTTTCGTACCACCTTGAGTCCAAAGAGACTAACGTTTAAGTGTAGGAAAGTACAGAGGTACTTTTTATCTTTCCTATTTAACCCGAGATAAGATTGTTCTCTTTGTGTTAAAAGGCTGTCAATTTACTTATTAAGAATTTGGAGTTTTTATTTTGCATCGTTTTGGAGTTAAGTTCTCATCTGTATTTCCAACACCGTACCCATCTTCCCTGAATACAGATCACTGTATCTAAACGATACGTATAGACTGGTTTTTAGGGGAGTACTAAAAGTTTAGTCTACGTATTGACGTAAATTACTCAAAGTGCTTAGTAATTATAGTCCGAGACTTTTGTCCTAGTAATTTTTACTAATTCTAAAAGTTGTTGTTTCTAAAATAGATTTATCATTCCTTTAATATAATCTACCGAGGTAAGGGACCCGTCAGATCAATTATTAAGTTGAACAAGTTTCGATATTTTATACCGTTGAATTTTTTAATACTTTCAAAAAGATTGAAAAAGAATTCCATCAACCGGTAAGAAGGAGAACCCTTATTCTTCCCGTCTAAAAGACAGCTTCGTAAAACGATAACGAAAAGTTAAAGGTAAATACCTTTTATCTATTATATCGTTAAAAAATCAATATTTAAATTTCAGTTTTACCTCTATTAATTCATAGAGGTAATTAATTGATTTCTAATTAAAAAGTTTTTTATCGTAAATGATTTAAGATTAGTTAGCTAACTGATAAAAATAGGGACCGAGTTACATTGAGAGAAACTGTATAAGAACAGAATTAAGATCAAATACAGTTGTTACACATTAGTTTCTACTGAGAAACGGTAAAACTAGAGAGTCAAGATAATTGTGGAAACTAGATTGCTATTACAAAAGGTTGTCTTAATAAAGATGAGTTTGAAGACTCTACTTTAAACACCTAAATAAGAGTGGGGTTAGTTAAAAAGTTGAGAGTCCTGTGGTTTACCACAGGATGTTAAGTTAAGTTAGGACTCTGATAGATAGACCTCTATCGTAGTCTAGGGTGTTCAAGTCCCAAGTCCGGGTGTTCTGACGGGGGGGGGGGTGGTAATCTCTCCTCAACACCCGGACAGTATGAGGACTGGTTGACCGATATTTAATCCTGAAGGTACCTAAGGAAAGGTCCAAGATATTAAACGATTCTACCGAGTGTCTTGGGTTCTTTCGTGAAACGAATGTCAATGACAGACTATGTTTAGTCCTTGACAGTTTATCTTCTCTGTGTATCCCGGTCCATACCACTTTTCCCCGTGTTTCGACGGTATACTGGAGATTCATATAGTAACATAGTGGGAGGGTCGTGGAGTCTCCACTACCTCCTCTCCCGACTGTCAAAATTGGTAGATCAATACGGAAGTACAAAGACCACAGTTCGTGGGTGGAACTTCGGTATACCCAGAGGTCGGTGGTGAGTACAGTACAGTTATCGTATTTGAGTCCATACCGATTCTCCCTGAATACTGATCATTTCCTACCAGGAGAGTAGAGATAGTCCTTTAAAGTTCAAAATCCTCGAGACATGACACTTGGCCCCTCCTTCTGGTTTATTCTGTATGTGATCTGGTCAACACGATTGATGACACACGGAGTTTTAACGGATGTCAACACAATGTTATTCTCTAACGACTCTGGACGATAACCTTAAAGTCTAAACCTTCTAGACCCCCACATTGGACTCTTATACGTCAAGATCATTCAAGGGTCCACTACGACAACGACGATCGTGTCCCTGGTGTGAAGGTCTCCCTGAGTCACTTCGTAGCAGTAATGAATATTATCGTTGAGGAGACATCAGTGTTAAAGTAAGTA	0.0	q5	IMPRECISE;SVTYPE=INS;SVLEN=4411;END=0;CIPOS=-0,0;CILEN=-35,35;RE=6;RNAMES=m84088_250402_230943_s1/142935135/ccs,m84088_250402_230943_s1/228853310/ccs,m84088_250402_230943_s1/56887202/ccs,m84088_250402_230943_s1/74388006/ccs,m84088_250402_230943_s1/149423853/ccs,m84088_250402_230943_s1/66389265/ccs;AF=0.0321	GT:DR:DV:PL:GQ	0/0:181:6:0,217,878:217

So as you can see END is 0.

[d-cameron]

1.6.1.2 "Telomeres are indicated by using positions 0 or N+1, where N is the length of the corresponding chromosome or contig."

putting the first base G is kind of misleading - I posit that using N would

be preferable. That is what the example in Section 5.4.5 does.

…

On Tue, Jun 17, 2025 at 9:58 AM Tom Conway ***@***.***> wrote: *drtconway* left a comment (zaeleus/noodles#343) <#343 (comment)> Here's a complete record: chr4_GL000008v2_random 0 cuteSV.INS.8444 G GGTACTTTTTGTTTCTCAGTGACTTGTGACTATTAACAAAAAGAGAGAGAGAGAGTAAAAAAGGAAAAAAACCTCTACTTCAGAGTGAGACAACGGGTCCGACCTCACGTTACCGTTGTAGAGACAAGTGACGTTGGAAGTGTAGGACCCAAGTCCGCTAAAAGGACGGAGTCGGAGGGTTCATCGACCCTAAGGTCCACGTAGGGCGGTACGGACCTATTCAAAAACATAAAAGTCATCTCTGCCCCAAAGTAGTACAACCGGTCCGACCAGAGTTTGAGGACTGGAGTTTACTAGACGGGCGGAATCGGAGGGTTTAACGACCACTACTAAAAGTAAATTACCACAAGATGTTTCACTCAAATCCGTTACTTTCGCGTGTACAGTAGATATCAAATGGTCTTTCATCATGGCCAGTGGTCGTAACTTCGTATGAGGATTGTGACCTTTTAAAGAAGTCGTCGGTCTTAATTAGAGTCTAAAGCTCTGACCTCGTTTCTGATTAAACTATACACAAGTAAGTCTAATCATTTAGAGGTACCACAAACTTGACATATTATAGTTTAGATTTTATTCAGTTTTCTCAACTTGTTATAGAGTAGAGACTCTTTTTTTACGTAATAAAAGTTGTCTAGATTCTGACTTAAATTTTAGGTTGACGGATAAGGTCACAAACAAAGAGATCTTATTCCGGTTCGGGATGCTTAAGACCTATCACAGGAGGTTCAACATAAATTCACTATTAACTATTTAGATACCAAATCTATGAATACCTTAAGGTATATTGAGATCGAGGTCGGAAACACTCATTTTTAACCGTCACTAGTACCGGAATTACTCCCGTACACCAATTAACCTTAAAAGTCATGGTTTTCAGCAGAAGAATGATCAAATAGTCCCATGGGATATTATCATAGGTATTGAGTAACGGGTGAACAGACAAAACTAAAGAAAAAATTTTTCTGTTAAAATTTGAATCTCCTAAACATGGAGTTCAGTGTATACACCGTGTAAATATTTTATTAGTACGGAAACGTATAAAAAGGAAACTATTGAATCGATATATAGGTTCGGTAAATACAAAAAATTAACAGTGTTTATAACGTGACTTTATCTCGTGATCAACGTTTTTACGGACCACATTGTAAAGAAGAACAAGAACTTTCAATAAAATTATGGGGTACAAATTCTATAAGGATCTGTTGATATACATTGATATTAATAAAGACGATCCATCTTTGTAGAAACACCAGTTGTAAACACTTCACGGAACATTTATTTTTATACGAACTCAACTTTATCATAAAACGTTTATATCGGGATTTATCTCGATGACTAAAATGGTTCTTAATTTCTGAGGTTTAAGACAGAAACAGAAAGTTTAATCCAAATATCCTTGTAGAGTAATGTTACTCGGTTTTCGGATAGACTAGTAATTACTCTATTAAAACTGTGAAGGTTTACTGTGTCAATTAAACATCATCCAAAAACTTCACGTTTAATGTACAAAGACAAATTTATTACTGTCATTTTATACCTGTGATAGATAAGTGTGAGATATATGTGTTGATGTTGAACAGAGATCTTCTTTTCAAATTAAAAAGTCCAGAGTAGGTGTTTTACTTCACAATTATTCGACATAGAAATAAGAGTAAAAATTCATTAAAGAAGAATTGGTAACATATTGTGATGGTGTCAAACATACCTCTATTTTATTGAGTACCACTCTTTAGGTCAAGAAATTTTTAGTAAAATGGAAGTTAGTGAAAATCCTATAGATACCAACATCTCATTCCTACTTTTAAATATTTGTAATCAATTTAATACGTTGACAAAAACTTCTACAAAGTTTTAATGTAAAAGGAATAAAATGTTGATCAAAAGTACATAAGTGTTTACTCTGTGTACTCTTAATAATTACGTTTTCGTGTCGCATAGTTGTTTCAAATATTACATTTATCAAATCAATCCTTTCTTACATTACTCACTTAATAGTTTCGATAAAATAACCAAGGATCTAAATGGTATCTTCTTTTATTTCGTCGGAATATTTTTAAAGAGTTGTGACTCAGGTAAATTCCAGATAAAACATACGGCATTCAAGATTTTTTACAATAACGGCCATTTTTATCCTACGGATAGATTTTAACAGAAAAATTTCCAAGAAAGAATAAGAAGAAGGAAAGGATACACACGTGTCTTATGACAGTAGTAAAACTAACCTTGTGACAGAGAGAAAGAGTCTCCCATATGTAAACATCTTCCCTCTTTTATAAACAATTAAAACGAAGGATATTCTATTCCAGAGTTACCAACTTCCAGTCAATTGTTTGTAAAAAAAAACAGAAGTATGATAATAGAACCTTAATGGTTGACCGAAGATAAATCTAAAACTTAATACTTTCTTTTCTTGATCGTAAACAATCTTGAATGACACAAGGTCCTTGAAATAACTATGGTAAAGAAGAGTAATGGTGTAAACAAGAAAGTTAAGCAGAATAAAAAAAGGTAGAATTATTCCACCCTGGTCTCCAAAACATTATAGGTGATTGGGTGACACCTAAAGTACCGAACGTTCTCCTTCTTTCGTACCACCTTGAGTCCAAAGAGACTAACGTTTAAGTGTAGGAAAGTACAGAGGTACTTTTTATCTTTCCTATTTAACCCGAGATAAGATTGTTCTCTTTGTGTTAAAAGGCTGTCAATTTACTTATTAAGAATTTGGAGTTTTTATTTTGCATCGTTTTGGAGTTAAGTTCTCATCTGTATTTCCAACACCGTACCCATCTTCCCTGAATACAGATCACTGTATCTAAACGATACGTATAGACTGGTTTTTAGGGGAGTACTAAAAGTTTAGTCTACGTATTGACGTAAATTACTCAAAGTGCTTAGTAATTATAGTCCGAGACTTTTGTCCTAGTAATTTTTACTAATTCTAAAAGTTGTTGTTTCTAAAATAGATTTATCATTCCTTTAATATAATCTACCGAGGTAAGGGACCCGTCAGATCAATTATTAAGTTGAACAAGTTTCGATATTTTATACCGTTGAATTTTTTAATACTTTCAAAAAGATTGAAAAAGAATTCCATCAACCGGTAAGAAGGAGAACCCTTATTCTTCCCGTCTAAAAGACAGCTTCGTAAAACGATAACGAAAAGTTAAAGGTAAATACCTTTTATCTATTATATCGTTAAAAAATCAATATTTAAATTTCAGTTTTACCTCTATTAATTCATAGAGGTAATTAATTGATTTCTAATTAAAAAGTTTTTTATCGTAAATGATTTAAGATTAGTTAGCTAACTGATAAAAATAGGGACCGAGTTACATTGAGAGAAACTGTATAAGAACAGAATTAAGATCAAATACAGTTGTTACACATTAGTTTCTACTGAGAAACGGTAAAACTAGAGAGTCAAGATAATTGTGGAAACTAGATTGCTATTACAAAAGGTTGTCTTAATAAAGATGAGTTTGAAGACTCTACTTTAAACACCTAAATAAGAGTGGGGTTAGTTAAAAAGTTGAGAGTCCTGTGGTTTACCACAGGATGTTAAGTTAAGTTAGGACTCTGATAGATAGACCTCTATCGTAGTCTAGGGTGTTCAAGTCCCAAGTCCGGGTGTTCTGACGGGGGGGGGGGTGGTAATCTCTCCTCAACACCCGGACAGTATGAGGACTGGTTGACCGATATTTAATCCTGAAGGTACCTAAGGAAAGGTCCAAGATATTAAACGATTCTACCGAGTGTCTTGGGTTCTTTCGTGAAACGAATGTCAATGACAGACTATGTTTAGTCCTTGACAGTTTATCTTCTCTGTGTATCCCGGTCCATACCACTTTTCCCCGTGTTTCGACGGTATACTGGAGATTCATATAGTAACATAGTGGGAGGGTCGTGGAGTCTCCACTACCTCCTCTCCCGACTGTCAAAATTGGTAGATCAATACGGAAGTACAAAGACCACAGTTCGTGGGTGGAACTTCGGTATACCCAGAGGTCGGTGGTGAGTACAGTACAGTTATCGTATTTGAGTCCATACCGATTCTCCCTGAATACTGATCATTTCCTACCAGGAGAGTAGAGATAGTCCTTTAAAGTTCAAAATCCTCGAGACATGACACTTGGCCCCTCCTTCTGGTTTATTCTGTATGTGATCTGGTCAACACGATTGATGACACACGGAGTTTTAACGGATGTCAACACAATGTTATTCTCTAACGACTCTGGACGATAACCTTAAAGTCTAAACCTTCTAGACCCCCACATTGGACTCTTATACGTCAAGATCATTCAAGGGTCCACTACGACAACGACGATCGTGTCCCTGGTGTGAAGGTCTCCCTGAGTCACTTCGTAGCAGTAATGAATATTATCGTTGAGGAGACATCAGTGTTAAAGTAAGTA 0.0 q5 IMPRECISE;SVTYPE=INS;SVLEN=4411;END=0;CIPOS=-0,0;CILEN=-35,35;RE=6;RNAMES=m84088_250402_230943_s1/142935135/ccs,m84088_250402_230943_s1/228853310/ccs,m84088_250402_230943_s1/56887202/ccs,m84088_250402_230943_s1/74388006/ccs,m84088_250402_230943_s1/149423853/ccs,m84088_250402_230943_s1/66389265/ccs;AF=0.0321 GT:DR:DV:PL:GQ 0/0:181:6:0,217,878:217 So as you can see END is 0. — Reply to this email directly, view it on GitHub <#343 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/ABOBYOGSERQ7ACDRDIITDNT3D5KYXAVCNFSM6AAAAAB7L3CGZWVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDSNZYGUYDKNBRGM> . You are receiving this because you are subscribed to this thread.Message ID: ***@***.***>

[zaeleus]

noodles is parsing the example record correctly. The variant caller is writing invalid data (END=0), and an error is subsequently thrown because it is a necessary, semantically relevant field in many use-cases (indexing, converting to BCF, etc.).

Since it is a well-formed record, noodles does successfully read it. A workaround would be to remove the invalid field before any calls that require resolving the variant span or end, e.g.,

for result in reader.records(&header) {
    let mut record = result
        .and_then(|r| vcf::variant::RecordBuf::try_from_variant_record(&header, r.as_ref()))?;

    info_remove_end_eq_0(record.info_mut());

    // ...
}

fn info_remove_end_eq_0(info: &mut vcf::variant::record_buf::Info) {
    use noodles_vcf::variant::{record::info::field::key, record_buf::info::field::Value};

    if let Some(Some(Value::Integer(0))) = info.get(key::END_POSITION) {
        info.as_mut().swap_remove(key::END_POSITION);
    }
}

Of course, this can be more stringent depending on the expected input data (values < 0, allowing non-numeric values, etc.).

---

If rec.variant_start() returns None for this case, would it be more consistent for rec.variant_end() to do the same here?

Variant records always have a span and end, even telomeric breakends. Missing positions have a virtual position of 1 because it's required by both indexing and calculation of the BCF rlen field.