4.11 KB
Newer Older
Yoann Pradat's avatar
Yoann Pradat committed
1 2 3 4 5 6 7 8 9 10 11 12
# Biotool for annotating variants from a VCF file.

The tool is divided in 3 steps
- Manual parsing of the VCF
- Run [vcf2maf]( to extract standard information 
- Run [Variant Ensembl's Predictor]( for annotations. 

## 1. What is the tool doing ?

VEP annotates variants with information from multiple external databases and can be configure for to answer a lot of specific needs. For more details,
see [VEP's options page]( VEP does not however extract information like number of reads or somatic status from the VCF file. vcf2maf is supposed to perform these tasks but failed to do on quite a lot of example VCF and does not always extract all the relevant information. For this reason, manual parsing was implemented in Python and was tested on TCGA VCF files from the legacy archive portal (see examples) and other VCFs. See the table for exhaustive details.

Pradat Yoann's avatar
Pradat Yoann committed
13 14
### 1.1. Manual parsing

Relies on tags specified by the user to extract relevant info like genotype (GT), somatic status (SS), quality and filter info (QUAL, INFO) and most importantly reads information (AD, DP, FA, DP4, TAR, TIR). The parser has been tested on VCF files as produced by
Pradat Yoann's avatar
Pradat Yoann committed
16 17 18 19 20
- Mutect v.1 (TCGA GA SNV) and Strelka (TCGA GA Indel)
- sets of callers VarScanSomatic-Strelka-Sniper-Samtools (TCGA HS SNP), GatkSomaticIndel-Pindel-Strelka-VarScanSomatic (TCGA HS Indel).
- Mutect v.1.1.7 with no header
- Strelka v.2.9.2

Yoann Pradat's avatar
Yoann Pradat committed
21 22 23 24 25 26 27 28 29 30 31 32 33
### 1.2. vcf2maf

vcf2maf also runs VEP internally but performs extra work to build some annotation fields including the following
- *Hugo_Symbol* (can be replaced by VEP's *Symbol*)
- *Entrez_Gene_Id*
- *NCBI_Build*
- *Chromosome* (can be extracted from VEP)
- *Start_Position*, *End_Position*
- *Variant_Classification* (rewriting of VEP's *Consequence*)
- *Variant_Type*
- *Reference_Allele*
- *Tumor_Seq_Allele1* (can be discarded)
- *Tumor_Seq_Allele2* (can be discarded)
- *dbSNP_RS*
Yoann Pradat's avatar
Yoann Pradat committed
35 36 37 38 39
- *HGVS_p*
- *HGVS_c*
- *HGVS_Short*
- *all_effects*

Pradat Yoann's avatar
Pradat Yoann committed
40 41
that are not available from VEP's output.

Yoann Pradat's avatar
Yoann Pradat committed
42 43 44 45 46
### 1.3. VEP

Run the VEP annotator on the VCF file from a specific set of options. The options can be changed in the code if required.

## 2. How to run the tool ?
47 48 49 50 51 52 53 54 55 56 57 58 59

### 2.1 Install VEP

Clone VEP from the [official github]( Installation details are provided there. As specified, run

git clone
cd ensembl-vep

The perl script may return errors as missing dependencies or other. For instance,the error `Bio::Root::Version is not installed` may be solved by running `sudo cpanm Bio::Root::Version`. You may have more than one such library to install. Refer to the github for the details.

The installation from the perl script offers the choice to install cache files (most efficient use of vep) and FASTA files (to retrieve sequence data for HGVS notations) into `$HOME/.vep`. You may also install plugins for additional analyses. Download cache files for Homo Sapiens genome 100_GRCh37 (or newer). The total download size is about 12 GB of data so a stable and fast connection is required here. The uncompressed FASTA file (**DO NOT FORGET** to uncompress this file or VEP will fail) Homo_sapiens.GRCh37.75.dna.primary_assembly.fa.gz requires about 3.0 GB of storage.
61 62 63

### 2.2 Install vcf2maf

VEP is required by vcf2maf but you also need the commands from `samtools` and `htslib` available at []( Do the following
65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83

cd samtools-1.x
./configure --prefix=/where/to/install
make install
ln -s /where/to/install/bin/samtools /usr/local/bin

cd htslib −1.x
./configure −−prefix=/where/to/install
make install

ln −s /where/to/install/bin/htslib /usr/local/bin 
ln −s /where/to/install/bin/tabix /usr/local/bin
ln −s /where/to/install/bin/bgzip /usr/local/bin

You may replace `/usr/local/bin` with whatever path where you usually save binaries. Finish the installation of vcf2maf following the instructions given in the [github](