Introduction

nf-pcgr is a bioinformatics analysis pipeline for the functional annotation and translation of somatic SNVs/InDels and copy number abberations for precision cancer medicine using Personal Cancer Genome Reporter (PCGR). nf-pcgr offers germline SNVs/INDELS intepretation and annotation using Cancer Predisposition Sequencing Reporter (CPSR).

The workflow has been designed to accept outputs generated by nf-core/sarek:

Tool	Germline	Somatic tumor-normal	Somatic tumor-only
CNVkit		✔️	✔️
DeepVariant	✔️
FreeBayes	✔️	✔️	✔️
HaplotypeCaller	✔️
Mutect2		✔️	✔️
Strelka somatic indels		✔️
Strelka somatic snvs		✔️
Strelka variants	✔️		✔️

Variant consolidation

Somatic variants called by multiple tools are reformatted to match PCGR specifications making them easily searchable in the HTML ouput.

Tumor sample depth (TDP), allele frequency (TAF) and allelic depths for the ref and alt (ADT) are manually calculated and when applicable, applied to the normal sample (NDP, NAF, ADN):

HCC1395T_vs_HCC1395N.freebayes.vcf.gz
#CHROM  POS     ID      REF     ALT     QUAL    FILTER  INFO    FORMAT  HCC1395_HCC1395N        HCC1395_HCC1395T
chr1    1212740 .       A       C       3793.78 PASS       AB=0;ABP=0;AC=2;AF=0.5;AN=4;AO=126;CIGAR=1X;DP=271;DPB=271;DPRA=0.868966;EPP=5.49198;EPPR=13.9276;GTI=0;LEN=1;MEANALT=1;MQM=60;MQMR=60;NS=2;NUMALT=1;ODDS=86.3557;PAIRED=1;PAIREDR=1;PAO=0;PQA=0;PQR=0;PRO=0;QA=4420;QR=5374;RO=145;RPL=102;RPP=107.861;RPPR=111.21;RPR=24;RUN=1;SAF=49;SAP=16.5217;SAR=77;SRF=64;SRP=7.33827;SRR=81;TYPE=snp;technology.ILLUMINA=1       GT:AD:AO:DP:GQ:PL:QA:QR:RO      0/0:145,0:0:145:99:0,436,4837:0:5374:145        1/1:0,126:126:126:99:3979,379,0:4420:0:0

TDP=126;NDP=145;TAF=1;NAF=0;ADT=0,126;ADN=145,0;TAL=freebayes

HCC1395T_vs_HCC1395N.mutect2.vcf.gz
#CHROM  POS     ID      REF     ALT     QUAL    FILTER  INFO    FORMAT  HCC1395_HCC1395N        HCC1395_HCC1395T
chr1    1212740 .       A       C       .       PASS    AS_SB_TABLE=63,80|49,76;DP=282;ECNT=1;MBQ=20,20;MFRL=151,154;MMQ=60,60;MPOS=30;NALOD=1.94;NLOD=25.89;POPAF=6.00;TLOD=341.76     GT:AD:AF:DP:F1R2:F2R1:FAD:SB 0/0:143,0:0.011:143:36,0:36,0:86,0:63,80,0,0    0/1:0,125:0.988:125:0,28:0,37:0,78:0,0,49,76

TDP=125;NDP=143;TAF=0.988;NAF=0.011;ADT=0,125;ADN=143,0;TAL=mutect2

HCC1395T_vs_HCC1395N.strelka.somatic_snvs.vcf.gz
#CHROM  POS     ID      REF     ALT     QUAL    FILTER  INFO    FORMAT  NORMAL  TUMOR
chr1    1212740 .       A       C       .       PASS    DP=271;MQ=60.00;MQ0=0;NT=ref;QSS=790;QSS_NT=3070;ReadPosRankSum=0.00;SGT=AA->AC;SNVSB=0.00;SOMATIC;SomaticEVS=19.73;TQSS=1;TQSS_NT=1    DP:FDP:SDP:SUBDP:AU:CU:GU:TU 145:0:0:0:145,145:0,0:0,0:0,0   126:0:0:0:0,0:126,126:0,0:0,0

TDP=126;NDP=145;TAF=1;NAF=0;ADT=0,126;ADN=145,0;TAL=strelka

Finally, the maximum values for TAF, TDP, NAF, NDP, ADT, ADN are taken as outputs for the consolidate variant call. In addition, values present in the ID and QUAL column (i.e not '.') are reported if present in any of the original calls:

1       1212740 .       A       C       3793.8  PASS    NDP=145;NAF=0.011;TDP=126;TAF=1;TAL=freebayes,mutect2,strelka

Pipeline summary

Quick Start

1. Install Nextflow (>=22.10.4)

2. Install Docker or Singularity

3. Download and unpack the human assembly-specific data bundle (grch38 for test-data):

• grch37 data bundle - 20220203 (approx 20Gb)

• grch38 data bundle - 20220203 (approx 21Gb)

GENOME="grch38" # or "grch37"
BUNDLE_VERSION="20220203"
BUNDLE="pcgr.databundle.${GENOME}.${BUNDLE_VERSION}.tgz"

wget http://insilico.hpc.uio.no/pcgr/${BUNDLE}
gzip -dc ${BUNDLE} | tar xvf -

Pass the directory containing the uncompressed data/ directory to nf-pcgr using the --database parameter for both PCGR and CPSR.

3. Download the pipeline and test it on a minimal dataset with a single command:

nextflow pull BarryDigby/nf-pcgr
nextflow run BarryDigby/nf-pcgr -profile test,<docker/singularity> --database '<path to PCGR database>'

Re-run the command if you encounter a FileNotFoundError - the test data did not fully download prior to workflow execution.

Parameter documentation

Detailed descriptions of parameters can be found at parameters.md or by running nextflow run BarryDigby/nf-pcgr --help.

Usage

Input samplesheet

The workflow accepts as input a samplesheet.csv file containing the paths to SNV/InDel VCF files and CNVKit copy number abberation .cns files. We have efforted to mimick the samplesheet specifications of nf-core/sarek for ease of use:

Column	Description
patient	Designates the patient/subject; must be unique for each patient, but one patient can have multiple samples
status	Normal/tumor (0/1) status of sample
sample	Designates the sample ID; must be unique. A patient may have multiple samples e.g a paired tumor-normal, tumor-only.
vcf	Full path to VCF file(s)
cna	Full path to CNS file

An example of a valid samplesheet is given below:

patient,status,sample,vcf,cna
HCC1395,1,HCC1395T,HCC1395T_vs_HCC1395N.mutect2.vcf.gz,HCC1395T.cns
HCC1395,1,HCC1395T,HCC1395T_vs_HCC1395N.freebayes.vcf.gz,HCC1395T.cns
HCC1395,1,HCC1395T,HCC1395T_vs_HCC1395N.strelka.somatic_snvs.vcf.gz,HCC1395T.cns
HCC1395,1,HCC1395T,HCC1395T_vs_HCC1395N.strelka.somatic_indels.vcf.gz,HCC1395T.cns
HCC1395,0,HCC1395N,HCC1395N.deepvariant.vcf.gz,
HCC1395,0,HCC1395N,HCC1395N.freebayes.vcf.gz,
HCC1395,0,HCC1395N,HCC1395N.haplotypecaller.vcf.gz,
HCC1395,0,HCC1395N,HCC1395N.strelka.variants.vcf.gz,
HCC1396,1,HCC1396T,HCC1396T_vs_HCC1396N.mutect2.vcf.gz,
HCC1396,1,HCC1396T,HCC1396T_vs_HCC1396N.freebayes.vcf.gz,
HCC1396,1,HCC1396T,HCC1396T_vs_HCC1396N.strelka.somatic_snvs.vcf.gz,
HCC1396,1,HCC1396T,HCC1396T_vs_HCC1396N.strelka.somatic_indels.vcf.gz,

copy number abberation .cns files must be present for every sample entry when --cna_analysis true.

File names

Input VCF file names must contain a string denoting the variant calling tool used to detect variants between the first and second period character:

HCC1396T_vs_HCC1396N.freebayes.vcf.gz

This is the default naming convention of nf-core/sarek, thus if your VCF files originate from a different workflow you must add them prior to running nf-pcgr. Accepted strings are deepvariant, freebayes, haplotypecaller, mutect2.

For files generated by Strelka, the workflow will consider the text between the first and third period characters: strelka.variants, strelka.somatic_indels, strelka.somatic_snvs:

HCC1396T_vs_HCC1396N.strelka.somatic_indels.vcf.gz

Sample outputs

PCGR

CPSR

Credits

nf-core/pcgr was originally written by Barry Digby.

We thank the following people for their extensive assistance in the development of this pipeline:

• Nathan Thorpe

• Sam Minot

• Sigve Nakken

Contributions and Support

Please open an issue or reach out to me (Barry Digby) on the nf-core slack channel.

I am interested in adding compatability for additional variant calling tools and optimising the intake of large VCF files.

Citations

Cancer Predisposition Sequencing Reporter (CPSR): A flexible variant report engine for high-throughput germline screening in cancer Nakken S, Saveliev V, Hofmann O, Møller P, Myklebost O, Hovig E.

Int J Cancer. 2021 Dec 1;149(11):1955-1960. doi:10.1002/ijc.33749

Personal Cancer Genome Reporter: variant interpretation report for precision oncology Nakken S, Fournous G, Vodák D, Aasheim LB, Myklebost O, Hovig E.

Bioinformatics. 2018 May 15;34(10):1778-1780. doi: 10.1093/bioinformatics/btx817

Sarek: A portable workflow for whole-genome sequencing analysis of germline and somatic variants Garcia M, Juhos S, Larsson M, Olason PI, Martin M, Eisfeldt J, DiLorenzo S, Sandgren J, Díaz De Ståhl T, Ewels P, Wirta V, Nistér M, Käller M, Nystedt B.

F1000Res. 2020 Jan 29;9:63. doi: 10.12688/f1000research.16665.2

The nf-core framework for community-curated bioinformatics pipelines.

Philip Ewels, Alexander Peltzer, Sven Fillinger, Harshil Patel, Johannes Alneberg, Andreas Wilm, Maxime Ulysse Garcia, Paolo Di Tommaso & Sven Nahnsen.

Nat Biotechnol. 2020 Feb 13. doi: 10.1038/s41587-020-0439-x.