Builiding reference for annovar based on build37.1?

Question

[Sharon]

Hi Everyone

Have anyone used annotation_variation.pl by Annovar to build a reference from NCBI build 37.1. I need this version exactly to annotate some old versions data. When I use UCSC hg19, the annotations makes less sense I think.

perl annotate_variation.pl -buildver b37 -downdb \
        -webfrom ftp://ussd-ftp.illumina.com/Homo_sapiens/NCBI/build37.1  Homo_sapiens_NCBI_build37.1  humandb/

I tried several url, I usually get:

NOTICE: Downloading annotation database ftp://ussd-ftp.illumina.com/Homo_sapiens/NCBI/build37.1/Homo_sapiens_NCBI_build37.1.txt.gz ... Failed WARNING: Some files cannot be downloaded, including ftp://ussd-ftp.illumina.com/Homo_sapiens/NCBI/build37.1/Homo_sapiens_NCBI_build37.1.txt.gz

I think it assumes the URL has the data file ends with .txt.gz. Any hint how to fix this? Having the full link doesn't work either

Thanks

[GenoMax]

Can you independently download the files first? Looks like -webfrom option is only to download from annovar servers.

[Sharon]

-webfrom allows from URL too as the documentation . But if I download it, how can I use what I downloaded in the command?

[igor]

I am not sure ANNOVAR can download from arbitrary URLs. It needs files to be in a specific format.

Answer 1

[igor]

It would probably be a lot easier to convert your VCF to hg19 than to prepare the ANNOVAR input for your reference, especially if you want to use more than just gene info (ANNOVAR has a lot of other pre-built databases). See this earlier discussion: build37 against hg19

[Sharon]

I will try that conversion. Thanks a lot igor.