James Bonfield

Unfortunately repurposing an existing option is trashing the intention of the original option. (Edit: updated with entirety of chr1.) Chr1 on SynDip with `--indel-bias 1.0` (the default) gives: ``` SNP...

My big concern with this is whether it'll actually last more than a few months as a half-solution. It's not working right now on all the data files, so we...

There are some examples in the man page here: http://www.htslib.org/doc/samtools.html#ENVIRONMENT_VARIABLES Also see the misc/seq_cache_populate.pl perl script for taking a fasta file and generating a series of nested directories named after...

Also you may consider using a caching proxy to the EBI so that ref sequences are locally cached and subsequent tools operate rapidly. @daviesrob has an example configuration to do...

By default samtools will cache (~/.cache/hts-ref) if you don't define REF_PATH/REF_CACHE, but it can eat into home directory space quickly so adjusting the defaults may be good. It also won't...

It's certainly possible to put other servers in there, including your own local ones, but I'm not aware of any public alternatives to the EBI one - yet! It's being...

I'd also argue that we need bcf-comaptible copies of these files. The lack of INFO and Contig lines means it's currently not possible to do round-trip tests. We should have...

Looking at genuine VCFs, I see 1000 genomes data containing: ``` ##reference=file:///gpfs/commons/datasets/old-nygc-resources/GRCh38_1000genomes/GRCh38_full_analysis_set_plus_decoy_hla.fa ``` So maybe we've been accepting for years VCF headers that aren't key=value. In that case the specification...

Oh, so when it says key = value it means a single `##key=value` - singular value, and not `##name=`? I assumed as it said "pairs" (pural) it was the latter,...

So going back to the two entries I see that cause bcftools breakage, is it the case that any "