VT normalize error with inconsistent fasta file for clinvar20181028 GRCh37 build

Question

[vnttung.iseartclub]

I keep ran into trouble while trying to normalize a clinvar VCF file with vt program, hg19 build. Tried every latest hg19/v37 reference fasta but still having the same error:

[variant_manip.cpp:96 is_not_ref_consistent] reference bases not consistent: Y:555381-555381 A(REF) vs N(FASTA)
[normalize.cpp:209 normalize] Normalization not performed due to inconsistent reference sequences. (use -n or -m option to relax this)

Do you know anywhere to find the reference file clinvar used to build their latest GRCh37 VCF, or anyway to solve this problem? Much appreciated!

Answer 1

[finswimmer]

Hello vnttung.iseartclub ,

this position is located in the PAR-Region. This region is usually masked with N on the Y chromosome in the reference files used for alignment. The reasons for that are described in more details in this tutorial.

I wonder how clinVar can be sure that this variant is located on Y and not on X. Nevertheless you have two option:

1. Ignore variants that are located in the PAR region of the Y chromosome for normalization
2. Find a reference sequence where this region isn't masked. One way is describe in Which human reference genome should I use?

fin swimmer

PS: @ Bastien Hervé This time I added the link to wiki again ;)