generate-names.pl only indexes the first few hundred names in a VCF

[rbuels]

On 11/14/2013 01:50 PM, Cris Lawrence wrote:

Robert:

OK...I run --safeMode. I have genes and snps The snps come from
dbsnp_137_hg19.vcf. The file includes some 56M SNPs. So I expected
this run to take a very long time, especially in safeMode. I was
prepared to let it go overnight if need be. Maybe 10 min,
later. The
run completes with the following:

bin/generate-names.pl --safeMode --verbose
Tracks:
mysnps
genes_by_gff3
DNA
study4514_analysis972.bam_coverage
Sampled input stats:
avg record text bytes 36.2166198432648
name input records 3243
operation stream estimated count 209601609
tracksWithNames ARRAY(0x2713770)
record stream estimated count 29943087
total namerec bytes 244933
total input bytes 1084437421
operations made 50001
namerecs buffered 6763
namerecs converted to operations 6763
Using 20-bit hashing
Hash store cache size: 894 buckets
Removing existing contents of target dir data/names
Hash store cache size: 894 buckets
Hash store cache size: 894 buckets

When I try searching for snps. I get an autocomplete list but
it is
clear that it is incomplete. SNPs I find at random in the
track, do
not
appear in the list. If I click on one in the list, one of two
things
happen:

1. I go to that snp.. but the snp I'm given appears to be a
   chr1 snp
   (I haven't seen a higher chr #)
2. Nothing happens. The GO button grays but the completed rs#
   (even
   when taken from the list offered) doesn't take me anywhere.

It seems to me that only a tiny fraction of the SNPs file is being
processed. Any suggestions as to what might be going awry?

• Cris