fetch contigs and lengths

[Daniel E Cook]

Is there a way to fetch the list of chromosomes from the header of a VCF along with their lengths?

[Brent Pedersen]

there is not currently, but I'd be open to adding this or getting a PR to implement. there's a method in in cyvcf2:seqnames which checks the header first and then checks for an index and reads from there if it can. but it returns only the names. still, that can be a starting point.

vcf.contigs is probably a better name

[Brent Pedersen]

there's a seqlens too.

[Daniel E Cook]

Ok - so after digging in a bit it looks like something more involved than this will be required:

var contigs = v.header.get("contig", BCF_HEADER_TYPE.BCF_HL_CTG)

II don't know C - but I might be able to make this work. Are you suggesting the C code be imported from cyvcf2? I can take a stab at it.

[Brent Pedersen]

I will implement on monday. I just meant to use cyvcf2 as a guide for the C functions to use.

[Brent Pedersen]

I started this in master. v.contigs will give you a seq of Contig objects, each with a name attribute.
there is also a length attribute, but it is not set (yet).
I'll have a look at that this week.

[Daniel E Cook]

Wow thank you! Yeah no need to rush to do this or anything, just wanted it to add some additional functions to my own little command-line utility.

[Brent Pedersen]

this would probably be better solved by exposing more/better wrapping for the bcf_rec_t structs and the header in general. i don't see an obvious, clean solution so i will let this marinate for a bit.

[Daniel E Cook]

Ok that sounds good. I like how hash tables are returned with info columns but that doesn’t quite work with contigs because there are multiple... I’ll parse the header for now from the string into a list of hash tables.

[Brent Pedersen]

this is now implemented. if lengths are available in the header, they will be set correctly on the contigs that have been returned from vcf.contigs

[Daniel E Cook]

Thank you! I will make use of this in seq-collection

[Daniel E Cook]

test.vcf.gz

I am seeing -1 contig lengths:

Contig(name:"I", length:-1'i64)
Contig(name:"II", length:-1'i64)
Contig(name:"III", length:-1'i64)
Contig(name:"IV", length:-1'i64)
Contig(name:"V", length:-1'i64)
Contig(name:"X", length:-1'i64)
Contig(name:"MtDNA", length:-1'i64)

Even though the header looks like this:

##fileformat=VCFv4.2
##FILTER=<ID=PASS,Description="All filters passed">
##samtoolsVersion=1.1-4-g9c4099f+htslib-1.1-3-g5b98adc
##samtoolsCommand=samtools mpileup -t DP,DV,DP4,SP -g -f /lscr2/andersenlab/dec211/pyPipeline/genomes/WS245/c_elegans.PRJNA13758.WS245.genomic.fa.gz -r I:1-1000000 v2_snpset/bam/AB1.bam
##reference=file:///lscr2/andersenlab/dec211/pyPipeline/genomes/WS245/c_elegans.PRJNA13758.WS245.genomic.fa.gz
##contig=<ID=I,length=15072434>
##contig=<ID=II,length=15279421>
##contig=<ID=III,length=13783801>
##contig=<ID=IV,length=17493829>
##contig=<ID=V,length=20924180>
##contig=<ID=X,length=17718942>
##contig=<ID=MtDNA,length=13794>

I'll see if I can track down what is going on...

[Brent Pedersen]

make sure you have the latest version of hts-nim as the change was quite recent.

[Daniel E Cook]

Thank you very much I figured it out!