Having more extra annotation

[incardon]

In the attempt of having more extra annotations. Where for more I mean 50 - 60 columns. Scalability issues arise in the use of the Extra annos table.

A potential solution come from the use of an external http server (REST api) that feed json variant scores, annotations, or infos.

As base to assess performance I am using Vcf files (Vcf as term to indicate vc/bcf compressed /uncompressed indexed with tabix for random access). The choice of vcf files as main DB goes for the reason that SamTools/and other tools exists to work on such files to sort, normalize, annotate, validate, parse and in general work on variances. I tried to assess how fast an http server would be able to query a compressed vcf files in random positions using samtools in java.

Up to now for a standard bgz compressed vcf file I can query up to 200 variances a second. (single core). I did not tested using multiple cores, I am expecting linear scalability, not sure up to which limit.

In any case considering that one variant details unroll generate one query, and and export probably would generate around 200 queries, up to now numbers looks good to me

[holtgrewe]

@incardon thanks for the issue. Some questions:

• Could you adjust it to the template from the documentation?
• Could you elaborate the following a bit further?

  Scalability issues arise in the use of the Extra annos table.