OncoVI is a fully-automated Python implementation of the oncogenicity guidelines by Horak et al. (Genetics in Medicine, 2022).

Starting from the genomic location of the variants, OncoVI:

1. performs functional annotation based on the Variant Effect Predictor (VEP) from Ensembl;
2. collects biological evidences from the implemented publicly available resources;
3. classifies the oncogenicity of somatic variants, based on the point-based system for combining pieces of evidence defined by Horak et al.

More detailed information on the resources used by OncoVI, the implementation of the oncogenicity guidelines, and its application to real-world data can be found in our pre-print.

Workflow of OncoVI:

OncoVI was implemented and tested on a dedicated conda enviroment running on a remote server based on Ubuntu 20.04.4 long-term support (LTS) operating system. To use this repository make sure to have:

• conda >= 24.11.1
• python >=3.8.8

OncoVI has been tested on:

• Ensembl VEP conda package >=v.111
• VEP Plugins (dbNSFP and spliceAI)

OncoVI supports only variants in human genome assembly GRCh38. No other genome assemblies are currently supported.

Clone the GitHub repository:

git clone https://github.com/MGCarta/oncovi.git

1. Create the conda environment envpy310

conda create --name envpy310 python=3.10

2. Activate the conda environment

conda activate envpy310

3. Install pandas

pip install pandas

4. Install colorama

pip install colorama

5. Install openpyxl

pip install openpyxl

The download and preparation of the ClinVar resource utilised by the functional annotation STEP is handled by the script 01_clinvar_resource_manager.sh.

# Move to the folder where the script is located
cd oncovi/src/

# Run the bash script
bash 01_clinvar_resource_manager.sh

1. Install the conda package Ensembl Vep

conda install bioconda::ensembl-vep

2. Run the installer (the latest version of the package will be automatically installed)

vep_install --NO_HTSLIB -a fp -s homo_sapiens -y GRCh38 --PLUGINS all -c '../.vep' -r '../.vep/Plugins/'

3. Install the cache files

# Move to the folder generated to save cache files
cd ../.vep

# Download the cache files
wget ftp://ftp.ensembl.org/pub/release-114/variation/indexed_vep_cache/homo_sapiens_refseq_vep_114_GRCh38.tar.gz

# Extract the archive
tar -xzf homo_sapiens_refseq_vep_114_GRCh38.tar.gz

# Remove the archive
rm homo_sapiens_refseq_vep_114_GRCh38.tar.gz

The dbNSFP plugin is used by the the functional annotation STEP. VEP reports detailed information on how to set up the dbNSFP plugin here.

About dbNSFP data files:

• Download dbNSFP files from https://sites.google.com/site/jpopgen/dbNSFP.
• There are two distinct branches of the files provided for academic and commercial usage. Please use the appropriate files for your use case.
• The file must be processed depending on dbNSFP release version and assembly (see commands below).
• The resulting file must be indexed with tabix before use by this plugin (see commands below).

For release 4.9c:

version=4.9c
wget https://dbnsfp.s3.amazonaws.com/dbNSFP${version}.zip
unzip dbNSFP${version}.zip
zcat dbNSFP${version}_variant.chr1.gz | head -n1 > h

GRCh38/hg38 data

zgrep -h -v ^#chr dbNSFP${version}_variant.chr* | sort -k1,1 -k2,2n - | cat h - | bgzip -c > dbNSFP${version}_grch38.gz
tabix -s 1 -b 2 -e 2 dbNSFP${version}_grch38.gz

Replace in the script vep.sh of this repository the full path to the dbNSFP resource:

--plugin dbNSFP,/path/to/dbNSFP4.9c_grch38.gz

The spliceAI plugin is used during the the functional annotation STEP.

To download the data:

1. Go to https://basespace.illumina.com/s/otSPW8hnhaZR
2. Login with an Illumina account (a free account can be created if needed).
3. Accept the invitation to get access to the project Predicting splicing from primary sequence
4. Make sure that the Context to accept invite is set to Personal and click Accept. The project data will now be accessible in your account.
5. On the navigation bar at the top, click Projects and then click on Predicting splicing from primary sequence.
6. Below the project title, click Other datasets.
7. Click genome_scores_v1.3 in the list of files.
8. On the new page, click genome_scores_v1.3 again, which will expand to show the folder contents.
9. Click on the files (from those with raw in the filename) to download from the list: spliceai_scores.raw.indel.hg38.vcf.gz, spliceai_scores.raw.indel.hg38.vcf.gz.tbi, spliceai_scores.raw.snv.hg38.vcf.gz, spliceai_scores.raw.snv.hg38.vcf.gz.tbi.

Replace in the script vep.sh of this repository the full path to the spliceAI resources:

--plugin SpliceAI,snv=/path/to/spliceai_scores.raw.snv.hg38.vcf.gz,indel=/path/to/spliceaiceai_scores.raw.indel.hg38.vcf.gz,cutoff=0.5

Variants in both text format and in variant call format (VCF) are accepted by OncoVI. No other formats are currently supported.

An exemplary input data in text format is available under:

# ~/oncovi/testdata/SOP_table_union.txt

An exemplary input data in VCF format is available under:

# ~/oncovi/testdata/onco_som_GRCh38_ClinVar_vep_input.vcf

# Move to the directory where the python script 02_VEP_based_pipeline.py is located
cd ../src

# Run OncoVI
python 02_VEP_based_pipeline.py -i ../testdata/SOP_table_union.txt

The current output format produced by OncoVI is a tab-delimited excel file. It includes all the variants provided in input for which the functional annotation STEP and the OncoVI prediction STEP were successful. An exemplary output file is available here for inspection:

# ~/oncovi/exampleoutput/SOP_table_union_OncoVI_prediction.xlsx

The output file includes:

• Features characterising the variants that were collected during either the VEP annotation step (i.e., SYMBOL, HGVSc, HGVSp, MANE_SELECT, EXON, CHROM, POS, REF, ALT, Amino_acids, Existing_variation, Consequence, Transcript #1, gnomADe_AF, gnomADg_AF) or the ClinVar interrogation step (i.e., ClinVar_germline, ClinVar_germline_ReviewStatus, ClinVar_Oncogenicity, ClinVar_Oncogenicity_ReviewStatus, ClinVar_Somatic_Clinical_Impact, ClinVar_Somatic_ReviewStatus, Clinvar_url)

• Features calculated by OncoVI: the variant-specific score, i.e., the sum of the points associated with the criteria triggered by OncoVI (Points), the oncogenicity classification associated with the variant-specific score according to the point-based system provided by Horak et al. (Classification), the criteria triggered by OncoVI based on the evidences collected (Criteria)

Please, help us to improve OncoVI by describing your bug/issue in detail

The MIT license file applies to only the scripts within this repository.

OncoVI makes use of resources that require a dedicated license agreement, such as OncoKB and COSMIC. For these reasons, OncoVI is intended for research purposes only and its use outside of this context is under the responsibility of the user. Please, visit the relative websites and verify that you are part of an academic institution to freely use OncoVI. It is the user's responsibility to carefully check and comply with the licenses of the resources that need to be additionally installed to use OncoVI.

Please cite our preprint Oncogenicity Variant Interpreter (OncoVI): oncogenicity guidelines implementation to support somatic variants interpretation in precision oncology if you decide to use OncoVI.