parseVEP

Usage

usage: moPepGen parseVEP [-h] -i ['<files>'] [['<files>'] ...] -o <file>
                         --source SOURCE [-g <file>] [-a <file>]
                         [--reference-source {GENCODE,ENSEMBL}]
                         [--index-dir [<file>]] [--debug-level <value|number>]
                         [-q]

Parse VEP output tsv to the GVF format of variant records for moPepGen to call
variant peptides. The genome assembly FASTA and annotation GTF must come from
the same GENCODE/ENSEMBL version, and must the consistent with the VEP output.

optional arguments:
  -h, --help            show this help message and exit
  -i ['<files>'] [['<files>'] ...], --input-path ['<files>'] [['<files>'] ...]
                        File path to the VEP output TXT file. Can take
                        multiple files. Valid formats: ['.tsv', '.txt',
                        '.tsv.gz', '.txt.gz'] (default: None)
  -o <file>, --output-path <file>
                        File path to the output file. Valid formats: ['.gvf']
                        (default: None)
  --source SOURCE       Variant source (e.g. gSNP, sSNV, Fusion) (default:
                        None)
  --debug-level <value|number>
                        Debug level. (default: INFO)
  -q, --quiet           Quiet (default: False)

Reference Files:
  -g <file>, --genome-fasta <file>
                        Path to the genome assembly FASTA file. Only ENSEMBL
                        and GENCODE are supported. Its version must be the
                        same as the annotation GTF and proteome FASTA
                        (default: None)
  -a <file>, --annotation-gtf <file>
                        Path to the annotation GTF file. Only ENSEMBL and
                        GENCODE are supported. Its version must be the same as
                        the genome and proteome FASTA. (default: None)
  --reference-source {GENCODE,ENSEMBL}
                        Source of reference genome and annotation. (default:
                        None)
  --index-dir [<file>]  Path to the directory of index files generated by
                        moPepGen generateIndex. If given, --genome-fasta,
                        --proteome-fasta and --anntotation-gtf will be
                        ignored. (default: None)

Arguments

-h, --help

show this help message and exit

-i, --input-path ['<files>'] Path

File path to the VEP output TXT file. Can take multiple files. Valid formats: ['.tsv', '.txt', '.tsv.gz', '.txt.gz']

-o, --output-path <file> Path

File path to the output file. Valid formats: ['.gvf']

--source str

Variant source (e.g. gSNP, sSNV, Fusion)

-g, --genome-fasta <file> Path

Path to the genome assembly FASTA file. Only ENSEMBL and GENCODE are supported. Its version must be the same as the annotation GTF and proteome FASTA

-a, --annotation-gtf <file> Path

Path to the annotation GTF file. Only ENSEMBL and GENCODE are supported. Its version must be the same as the genome and proteome FASTA.

--reference-source str

Source of reference genome and annotation.
Choices: ['GENCODE', 'ENSEMBL']

--index-dir <file> Path

Path to the directory of index files generated by moPepGen generateIndex. If given, --genome-fasta, --proteome-fasta and --anntotation-gtf will be ignored.

--debug-level <value|number> str

Debug level. str
Default: INFO

-q, --quiet

Quiet
Default: False

Running VEP for moPepGen

moPepGen relies on Ensembl Variant Effect Predictor (VEP) to annotate single nucleotide and small insertion and deletion (indel) variants, typically presented in VCF files. Note that VEP supports the annotation of multi-nucleotide variants and structural variants, but those are not currently supported by moPepGen and will be discarded by parseVEP.

Please refer to the official VEP Tutorial for instructions on downloading, installing and running VEP. The key is to ensure that the annotation version used in VEP is the same as the one you would like to use with moPepGen.

parseVEP currently only supports the TSV output of VEP, please set the suffix of --output_file to .tsv to ensure the correct format is outputted by VEP. VEP's default output format is tab-delimited, and the --tab option adds extra columns to the output that are not used by moPepGen. So including or excluding --tab will not impact the results of moPepGen.

Using an Ensembl GTF

VEP automatically uses an Ensembl GTF for annotation. Please ensure that the release version and genome build version is consistent with the references files downloaded. This is available in the name of the cache in output headers (for example, 107_GRCh38).

One can ensure that the correct version is used by downloading the cache following instructions and specifying the parameters: --species homo_sapiens --assembly GRCh38 --cache --cache_version 107

Using a Non-Ensembl GTF

If you have decided to use a set of reference files from GENCODE, it is important to supply VEP with the non-Ensembl GTF during annotation, so that chromosome names, transcript IDs and transcript coordinates match with your intended reference files. This can be specified in VEP.

As instructed by VEP, your GTF file must be sorted in chromosomal order and indexed.

grep -v "#" PATH_TO_GTF | sort -k1,1 -k4,4n -k5,5n -t$'\t' | bgzip -c > PATH_TO_GTF_GZ
tabix -p gff PATH_TO_GTF_GZ # gtf is not a tabix format option, gff works

To use the GTF for annotation, run VEP with the additional parameters

--custom PATH_TO_GTF,GENCODE,gtf --fasta PATH_TO_GENOME_FA

Followed by VEP, we run the filter_vep command to subset outputs to only those annotated using the GENCODE GTF, since VEP automatically outputs both native (Ensembl) and custom annotations.

filter_vep -i VEP_OUTPUT_TSV -o FILTERED_VEP_OUTPUT_TSV --filter Source = GENCODE

Tips for Running VEP

We recommend the following parameters for running VEP for run time optimization, please select appropriate settings for your system.

• --offline --cache
• --no_stats
• --fork
• --buffer_size
• --distance 0 - only annotation variants directly overlapping gene coordinates
• --chr 1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,X,Y,MT - focus on primary assembly
• --no_intergenic - ignore intergenic variants

We do not recommend adding any additional annotation flags since the information is not utilized by moPepGen and increases run time.