1 Prerequisites

1.1 Update from previous stable versions

PureCN is fully backward compatible with input generated by versions 1.10, 1.12 and 1.14, but 1.16 slightly changed the mapping bias database and incorporated the interval weights into the database directly. Simply re-create this RDS file to take advantage of the new features:

$ Rscript $PURECN/NormalDB.R --outdir $OUT_REF \
    --coveragefiles example_normal.list \
    --genome hg19 --normal_panel $NORMAL_PANEL --assay agilent_v6

PureCN 1.10 introduced a completely new normal database format with several important improvements such as not splitting the database by sample sex for the normalization of autosomes. It is therefore necessary (not only recommended) to re-run the NormalDB.R when upgrading from version 1.8.

For upgrades from version 1.6, we highly recommend starting from scratch following this tutorial.

1.2 Installation

For the command line scripts described in this tutorial, we will need to install PureCN with suggested dependencies:

if (!requireNamespace("BiocManager", quietly = TRUE))
    install.packages("BiocManager")
BiocManager::install("PureCN", dependencies = TRUE)

Alternatively, manually install the packages required by the command line scripts:

BiocManager::install(c("PureCN", "optparse", 
    "TxDb.Hsapiens.UCSC.hg19.knownGene", "org.Hs.eg.db"))

(Replace hg19 with your genome version).

To use the alternative and in many cases recommended PSCBS segmentation:

# default PSCBS without support of interval weights
BiocManager::install("PSCBS")

# patched PSCBS with support of interval weights
BiocManager::install("lima1/PSCBS", ref="add_dnacopy_weighting")

To call mutational signatures, install the GitHub version of the deconstructSigs package:

BiocManager::install("raerose01/deconstructSigs")

2 Prepare environment and assay-specific reference files

  • Start R and enter the following to get the path to the command line scripts:
system.file("extdata", package="PureCN")
## [1] "/tmp/RtmpW51WTJ/Rinstf72558d1df0/PureCN/extdata"
  • Exit R and store this path in an environment variable, for example in BASH:
$ export PURECN="/path/to/PureCN/extdata"
$ Rscript $PURECN/PureCN.R --help
Usage: /path/to/PureCN/inst/extdata/PureCN.R [options] ...
  • Generate an interval file from a BED file containing baits coordinates (not necessarily required with third-party segmentations, see in the corresponding Section 5):
# specify path where PureCN should store reference files
$ export OUT_REF="reference_files"
$ Rscript $PURECN/IntervalFile.R --infile baits_hg19.bed \ 
    --fasta hg19.fa --outfile $OUT_REF/baits_hg19_intervals.txt \
    --offtarget --genome hg19 \
    --export $OUT_REF/baits_optimized_hg19.bed \
    --mappability wgEncodeCrgMapabilityAlign100mer.bigWig \
    --reptiming wgEncodeUwRepliSeqK562WaveSignalRep1.bigWig

Internally, this script uses rtracklayer to parse the infile. Make sure that the file format matches the file extension. See the rtracklayer documentation for problems loading the file. Check that the genome version of the baits file matches the reference. Do not include chrM baits in case the capture kit includes some.

The --offtarget flag will include off-target reads. Including them is recommended except for Amplicon data.

The --genome version is needed to annotate exons with gene symbols. Use hg19/hg38 for human genomes, not b37/b38. You might get a warning that an annotation package is missing. For hg19, install TxDb.Hsapiens.UCSC.hg19.knownGene in R.

The --export argument is optional. If provided, this script will store the modified intervals as BED file for example (again every rtracklayer format is supported). This is useful when the coverages are calculated with third-party tools like GATK.

The --mappability argument should provide a rtracklayer parsable file with a mappability score in the first meta data column. If provided, off-target regions will be restricted to regions specified in this file. On-target regions with low mappability will be excluded. For hg19, download the file from the UCSC website. Choose the kmer size that best fits your average mapped read length. For hg38, download recommended 76-kmer or 100-kmer mappability files through the courtesy of the Waldron lab from:

See the FAQ section of the main vignette for instruction how to generate such a file for other references.

Similarly, the --reptiming argument takes a replication timing score in the same format. If provided, GC-normalized and log-transformed coverage is tested for a linear relationship with this score and normalized accordingly. This is optional and provides only a minor benefit for coverage normalization, but can identify samples with high proliferation. Requires --offtarget to be useful.

3 Create VCF files

PureCN does not ship with a variant caller. Use a third-party tool to generate a VCF for each sample.

Important recommendations:

  • Use MuTect 1.1.7 if possible

  • VCFs from most other tumor-only callers such as MuTect 2, VarScan2 and FreeBayes are supported, but only very limited artifact filtering will be performed for these callers. Make sure to provide filtered VCFs. See the FAQ section in the main vignette for common problems and questions related to input data.

  • Since germline SNPs are needed to infer allele-specific copy numbers, the provided VCF needs to contain both somatic and germline variants. Make sure that upstream filtering does not remove high quality SNPs, in particular due to presence in germline databases.

  • Run the variant caller with a 50-75 base pair interval padding to increase the number of heterozygous SNPs

4 Run PureCN with internal segmentation

The following describes PureCN runs with internal copy number normalization and segmentation.

4.1 Coverage

For each sample, tumor and normal, calculate GC-normalized coverages:

# Calculate and GC-normalize coverage from a BAM file 
$ Rscript $PURECN/Coverage.R --outdir $OUT/$SAMPLEID \ 
    --bam ${SAMPLEID}.bam \
    --intervals $OUT_REF/baits_hg19_intervals.txt

# GC-normalize coverage from a GATK DepthOfCoverage file
Rscript $PURECN/Coverage.R --outdir $OUT/$SAMPLEID \
    --coverage ${SAMPLEID}.coverage.sample_interval_summary \ 
    --intervals $OUT_REF/baits_hg19_intervals.txt

Similar to GATK, this script also takes a text file containing a list of BAM or coverage file names (one per line). The file extension must be .list:

# Calculate and GC-normalize coverage from a list of BAM files 
$ Rscript $PURECN/Coverage.R --outdir $OUT \ 
    --bam normals.list \
    --intervals $OUT_REF/baits_hg19_intervals.txt \
    --cores 4

Important recommendations:

  • Only provide --keepduplicates or --removemapq0 if you know what you are doing and always use the same command line arguments for tumor and the normals

4.2 NormalDB

To build a normal database for coverage normalization, copy the paths to all GC-normalized normal coverage files in a single text file, line-by-line:

ls -a normal*loess.txt.gz | cat > example_normal.list

# From already GC-normalized files
$ Rscript $PURECN/NormalDB.R --outdir $OUT_REF \
    --coveragefiles example_normal.list \
    --genome hg19 --assay agilent_v6

# When normal panel VCF is available (highly recommended for unmatched samples)
$ Rscript $PURECN/NormalDB.R --outdir $OUT_REF \
    --coveragefiles example_normal.list \
    --genome hg19 --normal_panel $NORMAL_PANEL --assay agilent_v6

Important recommendations:

  • Consider generating different databases when differences are significant, e.g.ย for samples with different read lengths or insert size distributions

  • In particular, do not mix normal data obtained with different capture kits (e.g. Agilent SureSelect v4 and v6)

  • Provide a normal panel VCF here to precompute mapping bias for faster runtimes. The only requirement for the VCF is an AD format field containing the number of reference and alt reads for all samples. See the example file $PURECN/normalpanel.vcf.gz.

  • For ideal results, examine the interval_weights.png file to find good off-target bin widths. You will need to re-run IntervalFile.R with the --offtargetwidth parameter and re-calculate the coverages.

  • The --assay argument is optional and is only used to add the provided assay name to all output files

  • A warning pointing to the likely use of a wrong baits file means that more than 5% of targets have close to 0 coverage in all normal samples. A BED file with the low coverage targets will be generated in --outdir. If for any reason there is no access to the correct file, it is recommended to re-run the IntervalFile.R command and provide this BED file with --exclude.

4.3 PureCN

Now that the assay-specific files are created and all coverages calculated, we run PureCN.R to normalize, segment and determine purity and ploidy:

mkdir $OUT/$SAMPLEID

# Without a matched normal (minimal test run)
$ Rscript $PURECN/PureCN.R --out $OUT/$SAMPLEID \
    --tumor $OUT/$SAMPLEID/${SAMPLEID}_coverage_loess.txt.gz \
    --sampleid $SAMPLEID \
    --vcf ${SAMPLEID}_mutect.vcf \
    --normaldb $OUT_REF/normalDB_hg19.rds \
    --intervals $OUT_REF/baits_hg19_intervals.txt \
    --genome hg19 

# Production pipeline run
$ Rscript $PURECN/PureCN.R --out $OUT/$SAMPLEID \
    --tumor $OUT/$SAMPLEID/${SAMPLEID}_coverage_loess.txt.gz \
    --sampleid $SAMPLEID \
    --vcf ${SAMPLEID}_mutect.vcf \
    --statsfile ${SAMPLEID}_mutect_stats.txt \
    --normaldb $OUT_REF/normalDB_hg19.rds \
    --mappingbiasfile $OUT_REF/mapping_bias_hg19.rds \
    --intervals $OUT_REF/baits_hg19_intervals.txt \
    --snpblacklist hg19_simpleRepeats.bed \
    --genome hg19 \
    --force --postoptimize --seed 123

# With a matched normal (test run; for production pipelines we recommend the
# unmatched workflow described above)
$ Rscript $PURECN/PureCN.R --out $OUT/$SAMPLEID \
    --tumor $OUT/$SAMPLEID/${SAMPLEID}_coverage_loess.txt.gz \
    --normal $OUT/$SAMPLEID/${SAMPLEID_NORMAL}_coverage_loess.txt.gz \
    --sampleid $SAMPLEID \
    --vcf ${SAMPLEID}_mutect.vcf \
    --normaldb $OUT_REF/normalDB_hg19.rds \
    --intervals $OUT_REF/baits_hg19_intervals.txt \
    --genome hg19

# Recreate output after manual curation of ${SAMPLEID}.csv
$ Rscript $PURECN/PureCN.R --rds $OUT/$SAMPLEID/${SAMPLEID}.rds

Important recommendations:

  • Even if matched normals are available, it is often better to use the normal database for coverage normalization. When a matched normal coverage is provided with --normal then the pool of normal coverage normalization and denoising steps are skipped!

  • Always provide the normal coverage database to ignore low quality regions in the segmentation and to increase the sensitivity for homozygous deletions in high purity samples.

  • The normal panel VCF file is useful for mapping bias correction and especially recommended without matched normals. See the FAQ of the main vignette how to generate this file. It is not essential for test runs.

  • The MuTect 1.1.7 stats file (the main output file besides the VCF) should be provided for better artifact filtering. If the VCF was generated by a pipeline that performs good artifact filtering, this file is not needed.

  • The --postoptimize flag defines that purity should be optimized using both variant allelic fractions and copy number instead of copy number only. This results in a significant runtime increase for whole-exome data.

  • If --out is a directory, it will use the sample id as file prefix for all output files. Otherwise PureCN will use --out as prefix.

  • The --parallel flag will enable the parallel fitting of local optima. See BiocParallel for details. This script will use the default backend.

  • --funsegmentation PSCBS will become the new default in 1.18. Support for interval weights currently requires a patch (see Section 1.2).

  • Defaults are well calibrated and should produce close to ideal results for most samples. A few common cases where changing defaults makes sense:

    • High purity and high quality: For cancer types with a high expected purity, such as ovarian cancer, AND when quality is expected to be very good (high coverage, young samples), --maxcopynumber 8

    • Small panels with high coverage: --padding 100 (or higher), requires running the variant caller with this padding or without interval file. Use the same settings for the panel of normals VCF so that SNPs in the flanking regions have reliable mapping bias estimates. The --maxhomozygousloss parameter might also need some adjustment for very small panels with large gaps around captured deletions.

    • Cell lines: Safely skip the search for low purity solutions in cell lines: --maxcopynumber 8, --minpurity 0.9, --maxpurity 0.99. Add --modelhomozygous to find regions of LOH in samples without normal contamination (do not provide this flag when matched normal data are available in the VCF).

    • cfDNA: --minpurity 0.1, --minaf 0.01 (or lower) and --error 0.0005 (or lower, when there is UMI-based error correction). Note that the estimated purity can be very wrong when the true purity is below 5-7%; these samples are usually flagged as non-aberrant.

    • Amplicon data: --model betabin (Amplicon data is not officially supported)

    • All assays: --maxsegments should set to a value so that with few exceptions only poor quality samples exceed this cutoff. For cancer types with high heterogenity, it is also recommended to increase --maxnonclonal to 0.3-0.4 (this will increase the runtime significantly for whole-exome data).

5 Run PureCN with third-party segmentation

5.1 General usage

If you already have a segmentation from third-party tools (for example CNVkit, GATK4, EXCAVATOR2). For a minimal test run:

Rscript $PURECN/PureCN.R --out $OUT/$SAMPLEID  \
    --sampleid $SAMPLEID \
    --segfile $OUT/$SAMPLEID/${SAMPLEID}.cnvkit.seg \
    --vcf ${SAMPLEID}_mutect.vcf \
    --intervals $OUT_REF/baits_hg19_intervals.txt \
    --genome hg19 

See the main vignette for more details and file formats.

6 Biomarkers

Dx.R provides copy number and mutation metrics commonly used as biomarkers, most importantly tumor mutational burden (TMB), chromosomal instability (CIN) and mutational signatures.

# Provide a BED file with callable regions, for examples obtained by
# GATK CallableLoci. Useful to calculate mutations per megabase and
# to exclude low quality regions.
grep CALLABLE ${SAMPLEID}_callable_status.bed > \ 
    ${SAMPLEID}_callable_status_filtered.bed

# Only count mutations in callable regions, also subtract what was ignored
# in PureCN.R via --snpblacklist, like simple repeats, from the mutation per 
# megabase calculation
# Also search for the COSMIC mutation signatures
# (http://cancer.sanger.ac.uk/cosmic/signatures)
Rscript $PureCN/Dx.R --out $OUT/$SAMPLEID/$SAMPLEID \
    --rds $OUT/SAMPLEID/${SAMPLEID}.rds \
    --callable ${SAMPLEID}_callable_status_filtered.bed \
    --exclude hg19_simpleRepeats.bed \
    --signatures \
    --signature_databases signatures.exome.cosmic.v3.may2019

# Restrict mutation burden calculation to coding sequences
Rscript $PureCN/FilterCallableLoci.R --genome hg19 \
    --infile ${SAMPLEID}_callable_status_filtered.bed \
    --outfile ${SAMPLEID}_callable_status_filtered_cds.bed \
    --exclude '^HLA'
 
Rscript $PureCN/Dx.R --out $OUT/$SAMPLEID/${SAMPLEID}_cds \
    --rds $OUT/SAMPLEID/${SAMPLEID}.rds \
    --callable ${SAMPLEID}_callable_status_filtered_cds.bed \
    --exclude hg19_simpleRepeats.bed

Important recommendations:

  • Run GATK CallableLoci with --minDepth N where N is roughly 20% of the mean target coverage of all samples.

7 Reference


Table 1: IntervalFile
Argument name Corresponding PureCN argument PureCN function
--fasta reference.file preprocessIntervals
--infile interval.file preprocessIntervals
--offtarget off.target preprocessIntervals
--targetwidth average.target.width preprocessIntervals
--mintargetwidth min.target.width preprocessIntervals
--smalltargets small.targets preprocessIntervals
--offtargetwidth average.off.target.width preprocessIntervals
--offtargetseqlevels off.target.seqlevels preprocessIntervals
--mappability mappability preprocessIntervals
--minmappability min.mappability preprocessIntervals
--reptiming reptiming preprocessIntervals
--reptimingwidth average.reptiming.width preprocessIntervals
--genome txdb, org annotateTargets
--outfile
--export rtracklayer::export
--version -v
--force -f
--help -h

Table 2: Coverage
Argument name Corresponding PureCN argument PureCN function
--bam bam.file calculateBamCoverageByInterval
--bai index.file calculateBamCoverageByInterval
--coverage coverage.file correctCoverageBias
--intervals interval.file correctCoverageBias
--method method correctCoverageBias
--keepduplicates keep.duplicates calculateBamCoverageByInterval
--removemapq0 mapqFilter ScanBamParam
--outdir
--cores Number of CPUs to use when multiple BAMs are provided
--parallel Use default BiocParallel backend when multiple BAMs are provided
--seed
--version -v
--force -f
--help -h

Table 3: NormalDB
Argument name Corresponding PureCN argument PureCN function
--coveragefiles normal.coverage.files createNormalDatabase
--normal_panel normal.panel.vcf.file calculateMappingBiasVcf
--assay -a Optional assay name Used in output file names.
--genome -g Optional genome version Used in output file names.
--outdir -o
--version -v
--force -f
--help -h

Table 4: PureCN
Argument name Corresponding PureCN argument PureCN function
--sampleid -i sampleid runAbsoluteCN
--normal normal.coverage.file runAbsoluteCN
--tumor tumor.coverage.file runAbsoluteCN
--vcf vcf.file runAbsoluteCN
--rds file.rds readCurationFile
--mappingbiasfile mapping.bias.file setMappingBiasVcf
--normaldb normalDB (serialized with saveRDS) calculateTangentNormal, filterTargets
--segfile seg.file runAbsoluteCN
--logratiofile log.ratio runAbsoluteCN
--sex sex runAbsoluteCN
--genome genome runAbsoluteCN
--intervals interval.file runAbsoluteCN
--statsfile stats.file filterVcfMuTect
--minaf af.range filterVcfBasic
--snpblacklist snp.blacklist filterVcfBasic
--error error runAbsoluteCN
--dbinfoflag DB.info.flag runAbsoluteCN
--popafinfofield POPAF.info.field runAbsoluteCN
--mincosmiccnt min.cosmic.cnt setPriorVcf
--funsegmentation fun.segmentation runAbsoluteCN
--alpha alpha segmentationCBS
--undosd undo.SD segmentationCBS
--maxsegments max.segments runAbsoluteCN
--minpurity test.purity runAbsoluteCN
--maxpurity test.purity runAbsoluteCN
--minploidy min.ploidy runAbsoluteCN
--maxploidy max.ploidy runAbsoluteCN
--maxcopynumber test.num.copy runAbsoluteCN
--postoptimize post.optimize runAbsoluteCN
--bootstrapn n bootstrapResults
--modelhomozygous model.homozygous runAbsoluteCN
--model model runAbsoluteCN
--logratiocalibration log.ratio.calibration runAbsoluteCN
--maxnonclonal max.non.clonal runAbsoluteCN
--maxhomozygousloss max.homozygous.loss runAbsoluteCN
--outvcf return.vcf predictSomatic
--out -o
--parallel BPPARAM runAbsoluteCN
--cores BPPARAM runAbsoluteCN
--seed
--version -v
--force -f
--help -h

Table 5: Dx
Argument name Corresponding PureCN argument PureCN function
--rds file.rds readCurationFile
--callable callable callMutationBurden
--exclude exclude callMutationBurden
--maxpriorsomatic max.prior.somatic callMutationBurden
--signatures deconstructSigs::whichSignatures
--signature_databases deconstructSigs::whichSignatures
--out
--version -v
--force -f
--help -h