VCF annotation with ExAC API.

A VCF annotation tool (prototype) using vcftools, ExAC REST API, and gencode. The output file is a tab-separated text file with the information following:

1. Type of variation
2. Depth of sequence coverage
3. Number of reads supporting the variant
4. Ratio of reads supporting alternative vs reference allele
5. gene inforamtion from ExAC.

Quick start

*Please first download gencode annotation from here and uncompress it as: data/gencode.v19.annotation.gtf *The following code runs this VCF annotation tool on a dataset in the repo.

cd vcf_anno/code/    
python var_anno_exac_yz.py ../data/Challenge_data_clean.vcf ../data/Challenge_data_clean.vcf.annotated.txt

How to use

Dependencies

1. Python3 and modules: pandas, requests, json.
2. vcftools, bedtools.

The main variant annotation tool:

1. The main script: code/var_anno_exac_yz.py
2. Supporting data:
   • A file ranking the deleterious effect of variants: data/VEP_variant_function_scores_Koscielny17NAR.cleaned.txt
   • Gencode gene annotation: Gencode v19 should be downloaded from here and uncompressed to: data/gencode.v19.annotation.gtf
3. Input data:
   • A sample VCF file: data/Challenge_data_clean.vcf
4. Output result:
   • The annotated variants after running the main script on the Input data. data/Challenge_data_clean.vcf.annotated.txt

A small tool/resource for ranking deleterious effect of variants

* A table for ranking deleterious effect of variants was found in [Koscielny et al., 2017, NAR, Supplementary Table 2.](https://academic.oup.com/nar/article/45/D1/D985/2605745#51199338)
* I have copied the table here: data/VEP_variant_function_scores_Koscielny17NAR.cleaned.txt 
* It is then cleaned it up using this script: lib/vep_variant_order_clean.py.
* The cleaned table generated is here, good to use: data/VEP_variant_function_scores_Koscielny17NAR.cleaned.txt 

A small tool to cleanup vcf header

* Some versions of vcftools will give Warnings for commas in header lines, while some versions accept them. A small tool here is to generate a new VCF header file to remove those comma.
* Running lib/vcf_header_clean.py on the original data will generate the new VCF header file lib/newheader.tmp.txt, and I have used bcftools to reheader the original file ../data/Challenge_data\ \(1\).vcf to make the Warning-free input: data/Challenge_data_clean.vcf.