The Personal Cancer Genome Reporter (PCGR) is a stand-alone software package for functional annotation and translation of individual cancer genomes for precision oncology. Currently, it interprets both somatic SNVs/InDels and copy number aberrations. The software extends basic gene and variant annotations from the Ensembl’s Variant Effect Predictor (VEP) with oncology-relevant, up-to-date annotations retrieved flexibly through vcfanno, and produces interactive HTML reports intended for clinical interpretation.

• May 22nd 2019: 0.8.1 release
  • Added Cancer_NOS.toml for unspecified tumor types
  • Minor bugfixing
• May 20th 2019: 0.8.0 release
  • Bundle update (VEP, CIViC, UniProt, CancerMine, dbNSFP, OpenTargets, DisGeNET, TCGA, ICGC-PCAWG)
  • New functionality
    • Ranking of variants in tiers 3-4/noncoding according to association scores from Open Targets Platform,(Carvalho-Silva et. al, NAR, 2019)
    • Mutational burden in the context of TCGA distributions
    • More extensive variant filtering options for tumor-only runs
    • Possibility to feed a panel-of-normals VCF to PCGR for filtering purposes
    • Possibility to add somatic CNA plot to report (provided as image file)
    • Pre-made configuration files pr. tumor type
    • Change pick order for primary transcript (VEP)
  • Massive upgrade of the Cancer Predisposition Sequencing Reporter
    • Choice between > 30 different virtual cancer predisposition gene panels
    • Improved variant classification according to ACMG criteria
    • Simplified report structure - organized according to pathogenicity levels
• Nov 27th 2018: 0.7.0 release
  • Bundle update and bug fixing (see CHANGELOG )
  • Reporting germline variants for cancer predisposition? Check out github.com/sigven/cpsr
• May 14th 2018: 0.6.2.1 release
  • CHANGELOG
• May 9th 2018: 0.6.2 release
  • Fixed various bugs reported by users (see CHANGELOG)
  • Data bundle update (ClinVar, KEGG, CIViC, UniProt, DiseaseOntology)
• May 2nd 2018: 0.6.1 release
  • Fixed bugs in tier assignment
• April 25th 2018: 0.6.0 release
  • Updated data sources
  • Enabling specification of tumor type of input sample
  • New tier system for classification of variants (ACMG-like)
  • VCF validation can be turned off
  • Tumor DP/AF presets
  • JSON dump of report content
  • GRCh38 support
  • Runs under Python3
• November 29th 2017: 0.5.3 release
  • Fixed bug with propagation of default options
• November 23rd 2017: 0.5.2 release
• November 15th 2017: 0.5.1 pre-release
  • Bug fixing (VCF validation)
• November 14th 2017: 0.5.0 pre-release
  • Updated version of VEP (v90)
  • Updated versions of ClinVar, Uniprot KB, CIViC, CBMDB
  • Removal of ExAC (replaced by gnomAD), removal of COSMIC due to licensing restrictions
  • Users can analyze samples run without matching control (i.e. tumor-only)
  • PCGR pipeline is now configured through a TOML-based configuration file
  • Bug fixes / general speed improvements
  • Work in progress: Export of report data through JSON

• Report for a breast tumor sample (TCGA)
• Report for a colon adenocarcinoma sample (TCGA)

IMPORTANT: If you use PCGR, please cite the publication:

Sigve Nakken, Ghislain Fournous, Daniel Vodák, Lars Birger Aaasheim, Ola Myklebost, and Eivind Hovig. Personal Cancer Genome Reporter: variant interpretation report for precision oncology (2017). Bioinformatics. 34(10):1778–1780. doi:10.1093/bioinformatics/btx817

• VEP - Variant Effect Predictor v96 (GENCODE v30/v19 as the gene reference dataset)
• CIViC - Clinical interpretations of variants in cancer (May 18th 2019)
• ClinVar - Database of variants with clinical significance (May 2019)
• DoCM - Database of curated mutations (v3.2, Apr 2016)
• CBMDB - Cancer Biomarkers database (Jan 17th 2018)
• DisGeNET - Database of gene-tumor type associations (v6.0, Jan 2019)
• Cancer Hotspots - Resource for statistically significant mutations in cancer (v2 - 2017)
• dBNSFP - Database of non-synonymous functional predictions (v4.0, May 2019)
• TCGA - somatic mutations discovered across 33 tumor type cohorts (The Cancer Genome Atlas, release 16, Mar 2019)
• UniProt/SwissProt KnowledgeBase - Resource on protein sequence and functional information (2019_04, Apr 2019)
• Pfam - Database of protein families and domains (v32, Sep 2018)
• DGIdb - Database of targeted cancer drugs (v3.0.2, Jan 2018)
• ChEMBL - Manually curated database of bioactive molecules (v25.1, Mar 2019)
• CancerMine - Literature-derived database of tumor suppressor genes/proto-oncogenes (v12, May 2019)

An installation of Python (version 3.6) is required to run PCGR. Check that Python is installed by typing python --version in your terminal window. In addition, a Python library for parsing configuration files encoded with TOML is needed. To install, simply run the following command:

pip install toml

1. Install the Docker engine on your preferred platform
   • installing Docker on Linux
   • installing Docker on Mac OS
   • NOTE: We have not yet been able to perform enough testing on the Windows platform, and we have received feedback that particular versions of Docker/Windows do not work with PCGR (an example being mounting of data volumes)
2. Test that Docker is running, e.g. by typing docker ps or docker images in the terminal window
3. Adjust the computing resources dedicated to the Docker, i.e.:
   • Memory: minimum 5GB
   • CPUs: minimum 4
   • How to - Mac OS X

a. Clone the PCGR GitHub repository (includes run script and folder with configuration files pr tumor type): git clone https://github.com/sigven/pcgr.git

b. Download and unpack the latest data bundles in the PCGR directory

• grch37 data bundle - 20190519 (approx 15Gb)
• grch38 data bundle - 20190519 (approx 16Gb)
• Unpacking: gzip -dc pcgr.databundle.grch37.YYYYMMDD.tgz | tar xvf -

c. Pull the PCGR Docker image (dev) from DockerHub (approx 5.1Gb):

• docker pull sigven/pcgr:dev (PCGR annotation engine)

a. Download and unpack the latest software release (0.8.1)

b. Download and unpack the assembly-specific data bundle in the PCGR directory

• grch37 data bundle - 20190519 (approx 15Gb)

• grch38 data bundle - 20190519 (approx 16Gb)

  • Unpacking: gzip -dc pcgr.databundle.grch37.YYYYMMDD.tgz | tar xvf -

  A data/ folder within the pcgr-X.X software folder should now have been produced

c. Pull the PCGR Docker image (0.8.1) from DockerHub (approx 5.2Gb):

• docker pull sigven/pcgr:0.8.1 (PCGR annotation engine)

The PCGR workflow accepts two types of input files:

• An unannotated, single-sample VCF file (>= v4.2) with called somatic variants (SNVs/InDels)
• A copy number segment file

PCGR can be run with either or both of the two input files present.

• We strongly recommend that the input VCF is compressed and indexed using bgzip and tabix
• If the input VCF contains multi-allelic sites, these will be subject to decomposition
• Variants used for reporting should be designated as 'PASS' in the VCF FILTER column

The tab-separated values file with copy number aberrations MUST contain the following four columns:

• Chromosome
• Start
• End
• Segment_Mean

Here, Chromosome, Start, and End denote the chromosomal segment, and Segment_Mean denotes the log(2) ratio for a particular segment, which is a common output of somatic copy number alteration callers. Note that coordinates must be one-based (i.e. chromosomes start at 1, not 0). Below shows the initial part of a copy number segment file that is formatted correctly according to PCGR's requirements:

Chromosome	Start	End	Segment_Mean
1 3218329 3550598 0.0024
1 3552451 4593614 0.1995
1 4593663 6433129 -1.0277

There are pre-made configuration files pr. tumor type in the conf folder, formatted using TOML. In the configuration file, the user may configure a number of options in the PCGR workflow, related to the following:

• Sequencing depth/allelic support thresholds
• MSI prediction
• Mutational signatures analysis
• Mutational burden analysis (e.g. target size of region subject to sequencing)
• VCF to MAF conversion
• Tumor-only analysis options
  • tick on/off various filtering schemes for exclusion of germline variants
• VEP/vcfanno options
• Log-ratio thresholds for gains/losses in CNA analysis

See here for more details about the exact usage of the configuration options.

A tumor sample report is generated by calling the Python script pcgr.py, which takes the following arguments and options:

usage: pcgr.py [options] <PCGR_DIR> <OUTPUT_DIR> <GENOME_ASSEMBLY> <CONFIG_FILE> <SAMPLE_ID>

Personal Cancer Genome Reporter (PCGR) workflow for clinical interpretation of
somatic nucleotide variants and copy number aberration segments

positional arguments:
pcgr_dir              PCGR base directory with accompanying data directory,
			    e.g. ~/pcgr-0.8.1
output_dir            Output directory
{grch37,grch38}       Genome assembly build: grch37 or grch38
configuration_file    PCGR configuration file (TOML format, in conf/ folder)
sample_id             Tumor sample/cancer genome identifier - prefix for
			    output files

optional arguments:
-h, --help            show this help message and exit
--input_vcf INPUT_VCF
			    VCF input file with somatic query variants
			    (SNVs/InDels). (default: None)
--input_cna INPUT_CNA
			    Somatic copy number alteration segments (tab-separated
			    values) (default: None)
--input_cna_plot INPUT_CNA_PLOT
			    Somatic copy number alteration plot (default: None)
--pon_vcf PON_VCF     VCF file with germline calls from Panel of Normals
			    (PON) - i.e. blacklist variants (default: None)
--tumor_purity TUMOR_PURITY
			    Estimated tumor purity (between 0 and 1) (default:
			    None)
--tumor_ploidy TUMOR_PLOIDY
			    Estimated tumor ploidy (default: None)
--force_overwrite     By default, the script will fail with an error if any
			    output file already exists. You can force the
			    overwrite of existing result files by using this flag
			    (default: False)
--version             show program's version number and exit
--basic               Run functional variant annotation on VCF through
			    VEP/vcfanno, omit other analyses (i.e. CNA, MSI,
			    report generation etc. (STEP 4) (default: False)
--no_vcf_validate    Skip validation of input VCF with Ensembl's vcf-
			   validator (default: False)
--docker-uid DOCKER_USER_ID
			    Docker user ID. Default is the host system user ID. If
			    you are experiencing permission errors, try setting
			    this up to root (`--docker-uid root`) (default: None)
--no-docker           Run the PCGR workflow in a non-Docker mode (see
			    install_no_docker/ folder for instructions (default:
			    False)

The examples folder contain input files from two tumor samples sequenced within TCGA (GRCh37 only). It also contains PCGR configuration files customized for these cases. A report for a colorectal tumor case can be generated by running the following command in your terminal window:

python pcgr.py --input_vcf ~/pcgr-0.8.1/examples/tumor_sample.COAD.vcf.gz --input_cna ~/pcgr-0.8.1/examples/tumor_sample.COAD.cna.tsv --tumor_purity 0.9 --tumor_ploidy 2.0 ~/pcgr-0.8.1 ~/pcgr-0.8.1/examples grch37 ~/pcgr-0.8.1/examples/examples_COAD.toml tumor_sample.COAD

This command will run the Docker-based PCGR workflow and produce the following output files in the examples folder:

1. tumor_sample.COAD.pcgr_acmg.grch37.html - An interactive HTML report for clinical interpretation
2. tumor_sample.COAD.pcgr_acmg.grch37.pass.vcf.gz - Bgzipped VCF file with rich set of annotations for precision oncology
3. tumor_sample.COAD.pcgr_acmg.grch37.pass.tsv.gz - Compressed vcf2tsv-converted file with rich set of annotations for precision oncology
4. tumor_sample.COAD.pcgr_acmg.grch37.snvs_indels.tiers.tsv - Tab-separated values file with variants organized according to tiers of functional relevance
5. tumor_sample.COAD.pcgr_acmg.grch37.json.gz - Compressed JSON dump of HTML report content
6. tumor_sample.COAD.pcgr_acmg.grch37.cna_segments.tsv.gz - Compressed tab-separated values file with annotations of gene transcripts that overlap with somatic copy number aberrations

sigven AT ifi.uio.no