extract_snps.py throwing assertion error about hisat2

[infphilo]

The script isn't compatible with VCF format. In a future release of HISAT2, I'll modify the script to support VCF format. BTW, it is necessary that we need the same chromosome names: GRCh37 has 1, 2, 3, ... while snpCommon.txt has chr1, chr2, chr3, ...

[vivekbhr]

@infphilo thanks for the reply.. I guess "chr" isn't the issue in this case since I am using the snpCommons.txt from here (http://ccb.jhu.edu/software/hisat2/downloads/evaluation/data/snpCommon_22.txt), where we don't have "chr" in chromosome names.

First two lines from GRCh37_ensembl_genome.fa :

>1 dna_sm:chromosome chromosome:GRCh37:1:1:249250621:1 REF
nnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnn

However I tried to find a workaround to get the information from vcf file (since it seems in my case the snpCommons.txt is not available anyway) . I got something like this using my script:

# ID               CHR   TYPE       POS        AltBase/Length/Sequence
rs235202311 1   deletion    3010567 6
rs220682121 1   single  3010582 A
rs233361213 1   single  3010583 G
rs249503176 1   single  3010601 G
rs219761216 1   single  3010629 A
rs235606595 1   single  3010634 G
rs235822761 1   insertion   3010661 G,GGTTTT,GTT,GGTT
rs262127828 1   single  3010704 G
rs226648868 1   single  3010724 A
rs248800471 1   single  3010736 A
rs265573862 1   single  3010797 C

Do you think this will be a valid input to build the graph index?

[roryk]

Thanks for the merge-- sorry to prematurely close this issue; this fixes one issue, but I think somewhere in hisat2-build when it is parsing the headers in the FASTA file it has a similar logic, which leads to the snps/splice sites/exons being ignored, so I think there is one more place it has to change.