Skip to content

crukci-bioinformatics/ampliconseq

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

ampliconseq

Variant calling pipeline for amplicon sequencing data.

Table of Contents


ampliconseq is an analysis pipeline for calling single nucleotide variants (SNVs) and indels in targeted amplicon sequencing data. Variants are called using GATK HaplotypeCaller, preferred for germline or clonal somatic mutations, especially in FFPE samples, or VarDict which can identify low allele fraction SNVs in circulating tumour DNA from plasma samples. In addition to caller-specific filters, the pipeline models the background substitution noise at each amplicon position to identify and filter SNV calls with very low allele fractions that are not distinguishable from noise. Alignment and target coverage metrics are computed and compiled into a QC report. Variants are annotated using Ensembl Variant Effect Predictor (VEP).

The ampliconseq pipeline is executed using the Nextflow scientic workflow system and all dependencies and tools are packaged in a Docker container that can be run either using Docker or Singularity. The inputs to the pipeline are BAM files containing sequence reads aligned to the reference genome.

The ampliconseq pipeline has the following features:

  • Choice of variant callers: GATK HaplotypeCaller and VarDict
  • Alignment and coverage QC report using metrics calculated by Picard CollectAlignmentSummaryMetrics and CollectTargetedPcrMetrics
  • Annotation of variants using Ensembl Variant Effect Predictor (VEP)
  • Support for overlapping amplicon targets by partitioning reads prior to variant calling
  • Support for calling and filtering low allele fraction SNVs, e.g. for circulating tumour DNA in plasma samples with allele fractions down to 0.1%, by fitting probability distributions to model background noise
  • Specific calling of known mutations
  • Assignment of confidence level based on whether a variant is called or filtered in each of a set of replicate libraries (usually duplicate libraries)
  • Minimal barrier to installation with the only requirements being a Java runtime, Nextflow, and either Docker or Singularity to run a container in which all other dependencies and tools are packaged
  • Scales easily from deployment on multi-core workstation to high-performance compute cluster or cloud with only a simple configuration change
  • Accompanying visualization tool for viewing and assessing SNV calls

The ampliconseq pipeline was developed by the Bioinformatics Core in collaboration with James Brenton's research group at the Cancer Research UK Cambridge Institute (CRUK CI).


  1. Install Nextflow (Java 8 or later required).

     curl -s https://get.nextflow.io | bash
    

    This creates a file named nextflow in the current directory. For convenience, move this to some location on your PATH. See the Nextflow documentation for more details on installing Nextflow.

  2. Download Ensembl VEP cache.

    This step can be skipped if the VEP cache for the relevant species and genome assembly is already installed or if variant annotation is not required.

     nextflow run crukci-bioinformatics/ampliconseq \
         -main-script download_vep_cache.nf \
         -with-singularity \
         --vepCacheDir /path_to/vep_cache \
         --vepSpecies homo_sapiens \
         --vepAssembly GRCh37
    

    This will download the ampliconseq pipeline from GitHub including the single step workflow for downloading the VEP cache (download_vep_cache.nf). It will also download the Docker container in which Ensembl VEP is installed from Docker Hub and from this build a Singularity container. Use -with-docker to use Docker instead of Singularity.

    Substitute the top-level VEP cache directory as required; note that this step will fail if the directory doesn't already exist.

    The VEP cache can be quite large (around 15G for homo sapiens) and downloading and unpacking the cache may take several minutes.

  3. Create a samples file (samples.txt) containing the sample identifier and BAM file for each library.

  4. Create an amplicon coordinates file (amplicons.csv).

  5. Create a configuration file (ampliconseq.config) specifying the sample sheet, the amplicon coordinates file, the reference genome, the VEP cache directory, the variant caller and various other parameters.

  6. Run the ampliconseq pipeline specifying the configuration file and execution profile.

     nextflow run crukci-bioinformatics/ampliconseq \
         -config ampliconseq.config \
         -with-singularity \
         -profile bigserver \
         -with-report ampliconseq_report.html \
         -with-timeline ampliconseq_timeline.html
    

The ampliconseq pipeline is downloaded and run using the Nextflow workflow system. Dependencies, including GATK, VarDict, Picard, Ensembl Variant Effect Predictor, R and various R packages, are packaged as a Docker container that can be run with either Docker or Singularity. The container is also downloaded by Nextflow. The only requirements are a recent version of Nextflow and either Docker or Singularity. Nextflow requires Java 8 or above and can be installed as shown in the Quickstart section above (see the Nextflow documentation for more details).

Using the latest stable release of ampliconseq is recommended. A specific version of ampliconseq can be installed using nextflow pull with the -revision (or -r) option:

nextflow pull crukci-bioinformatics/ampliconseq -r 1.0.0

When a specific version of ampliconseq is installed in this way the revision also needs to be specified when running the pipeline using nextflow run.

nextflow run crukci-bioinformatics/ampliconseq -r 1.0.0 -c ampliconseq.config

Run nextflow info to view details about the currently installed version.

nextflow info crukci-bioinformatics/ampliconseq

The latest snapshot of ampliconseq will be downloaded and run if no revision is specified using the -r or -revision command line option when running ampliconseq for the first time. Subsequent runs will use this snapshot version but Nextflow detects if there have been revisions to the pipeline since and displays a message such as the following:

NOTE: Your local project version looks outdated - a different revision is available in the remote repository [961d1d72a2]

Run the following command to update ampliconseq to the latest revision on the master branch:

nextflow pull crukci-bioinformatics/ampliconseq -r master

Dependencies, including GATK, VarDict, Ensembl VEP, R and various R packages, are packaged in a Docker container that will be downloaded automatically by Nextflow.

ampliconseq can be run without a container by installing the following tools and packages:

  • R 4.1.0 or above and the following packages:
    • tidyverse
    • optparse
    • fitdistrplus
    • nozzle.r1
    • base64
    • svglite
    • rsvg
    • ComplexHeatmap (from Bioconductor)
  • GATK 4.2.0.0 or above (includes the Picard tools used to calculate various metrics)
  • VarDict (Java version) 1.8.2 or above
  • Ensembl Variant Effect Predictor release 104 or later

These can be installed manually or, more straightforwardly, using Conda. The pipeline assumes that the executables, R, gatk, vardict-java and vep, are available on your PATH. The Docker container recipe (Dockerfile) uses Conda to install the dependencies and the Conda environment file, conda.yml, located in the GitHub repository within the docker subdirectory can be used to install these dependencies such that the pipeline can be run without using the container.

    conda env create -f conda.yml

Additionally, the pipeline contains a number of custom Java tools written using the HTSJDK library. These are available from the releases page on the GitHub repository; download and unpack the tarball file named ampliconseq-1.0.0-tools.tar.gz, substituting the version number as appropriate, and ensure that the bin subdirectory is available on the PATH.


The ampliconseq pipeline requires a sample sheet file, an amplicon coordinates file and, optionally, a configuration file in which the input files and parameter settings are specified.

The input files are aligned sequence BAM files, in which there is a single BAM file for each library and where each library contains amplified DNA for all amplicons within the panel. The reference genome sequence FASTA file to which the sequence reads were aligned must be specified in the configuration file or as a command line argument; this needs to be indexed and have an accompanying sequence dictionary.

The samples sheet provides details about each of the amplicon libraries. It can be either a tab-delimited (TSV) or comma-separated value (CSV) file. By default, the ampliconseq pipeline expects a file named samples.csv in the directory in which the pipeline is run, but this can be can be changed within the configuration file or using a command line argument (see below).

The sample sheet is expected to have the following columns.

Column Required Description
ID yes The library identifier or barcode
Sample yes The name or identifier of the sample from which the library was created
BAM no The BAM file name (where the directory can be specified as a configuration parameter) or path (can be a relative or absolute path)

Replicate libraries created from the same sample will share the same Sample name or identifier. This sample-based grouping of libraries is used in the pipeline when creating the variant call summary table, in which variants called within replicates of the same sample are gathered and reported together with a confidence level. Is is also used when identifying possible sample library mispairings as part of the QC report; libraries are clustered based on variant allele fractions and replicate libraries from the same sample are expected to cluster together.

An example sample sheet containing duplicate libraries for each of two samples is given below. This is a small snippet of a sample sheet; runs typically contain tens or hundreds of libraries.

ID                  Sample        BAM
SLX-12850.FLD0011   JBLAB-2493    FLD0011.bam
SLX-12850.FLD0012   JBLAB-2493    FLD0012.bam
SLX-12850.FLD0013   JBLAB-3401    FLD0013.bam
SLX-12850.FLD0014   JBLAB-3401    FLD0014.bam

If the sample sheet does not contain a BAM column the pipeline will assume that BAM files follow a file naming convention in which the ID is the prefix to which the '.bam' extension is added, e.g. SLX-12850.FLD0011.bam for the first library in the example sample sheet given above. There is a bamDir configuration parameter that can be set in order to avoid having to specify the full path for each BAM file in the sample sheet; it is prepended to the BAM file name given in the sample sheet or to the default file name based on the ID if the sample sheet does not contain a BAM column.

The amplicon coordinates file provides the start and end coordinates for each amplicon and the start and end position for the target region that excludes the primer sequences. By default, the ampliconseq pipeline expects a file named amplicons.csv in the directory in which the pipeline is run, but this can be changed within the configuration file or using a command line argument (see below).

The following columns are all required.

Column Description
ID The amplicon identifier
Chromosome The chromosome
AmpliconStart The start coordinate of the amplicon (includes primer sequence)
AmpliconEnd The end coordinate of the amplicon (includes primer sequence)
TargetStart The start coordinate of the target region (excludes primer)
TargetEnd The end coordinate of the target region (excludes primer)

The following snippet from an amplicon coordinates file contains a set of amplicons targeting the TP53 gene.

ID              Chromosome  AmpliconStart  AmpliconEnd  TargetStart  TargetEnd
TP53_D0008_001  chr17       7572903        7573031      7572924      7573010
TP53_D0008_002  chr17       7573904        7574019      7573922      7574000
TP53_D0008_003  chr17       7573975        7574077      7573997      7574049
TP53_D0008_004  chr17       7576789        7576917      7576812      7576898
TP53_D0008_005  chr17       7576873        7576961      7576895      7576936
TP53_D0008_006  chr17       7576996        7577115      7577015      7577097
TP53_D0008_007  chr17       7577074        7577182      7577094      7577157
TP53_D0008_008  chr17       7577434        7577528      7577453      7577506
TP53_D0008_009  chr17       7577484        7577612      7577503      7577590
TP53_D0008_010  chr17       7577561        7577667      7577587      7577649

The ampliconseq pipeline has a number of configuration parameters. Use the --help option to see usage instructions and details of each.

nextflow run crukci-bioinformatics/ampliconseq --help

The following parameters can be configured. These can be set either as command line options or using a configuration file.

parameter default value description
samples samples.csv CSV or TSV file giving the sample name and BAM file for each library (ID and Sample columns required, optional BAM column).
bamDir Directory in which BAM files are located; paths to BAM files specified in the sample sheet are relative to this directory or to the launch directory if not specified. Alternatively, this parameter can be left unset and full paths given in the BAM column within the samples file.
amplicons amplicons.csv CSV/TSV file containing amplicon coordinates (ID, Chromosome, AmpliconStart, AmpliconEnd, TargetStart, TargetEnd columns required).
specificVariants CSV/TSV file containing specific (or known) variants that are included in the summary regardless of whether these are called or not (Sample, Chromosome, Position, Ref, Alt columns required).
blacklistedVariants CSV/TSV file containing blacklisted variants that will be filtered (Chromosome, Position, Ref, Alt columns required).
referenceGenomeFasta /reference_data/GRCh37.fa FASTA file containing the reference genome sequence (must be indexed and have an accompanying sequence dictionary).
vepAnnotation false Annotate variants with Ensembl VEP.
vepCacheDir /reference_data/vep_cache Directory in which Ensembl VEP cache files are installed.
vepSpecies homo_sapiens The species name of the VEP annotation cache.
vepAssembly GRCh37 The genome assembly of the VEP annotation cache.
vepPickOneAnnotationPerVariant true Pick one annotation for each annotation using VEP --pick option.
outputDir Directory to which output files are written or the launch directory if not specified.
variantCaller VarDict The variant caller (VarDict or HaplotypeCaller).
minimumAlleleFraction 0.01 Lower allele fraction limit for detection of variants (for variant callers that provide this option only).
maximumReadsPerAlignmentStart 2500 Maximum number of reads to retain per alignment start position; reads above this threshold will be downsampled (specific to GATK HaplotypeCaller).
minimumMappingQualityForPileup 1 Minimum mapping quality of reads to include in the pileup, i.e. when computing depths and allele fractions.
minimumBaseQualityForPileup 10 Minimum base quality at a given locus for reads to include in the pileup, i.e. when computing depths and allele fractions.
minimumDepthForBackgroundNoise 100 Minimum depth of coverage at a given locus for a library to be included when computing background noise.
excludeHighestFractionForBackgroundNoise 0.1 Fraction of measurements with the highest allele fraction to exclude from fitting a distribution to the background noise (assumes these are not due to error/noise).
maximumAlleleFractionForBackgroundNoise 0.03 Maximum allele fraction to include in fitting a distribution to the background noise (assumes anything above this is not due to error/noise).
minimumNumberForFittingBackgroundNoise 10 Minimum number of libraries required to fit a background noise distribution.
chunkSizeForFittingBackgroundNoise 500000 Maximum number of pileup count rows to process in a chunk when fitting background noise distributions.
readChunkSizeForFittingBackgroundNoise 100000 Chunk size for reading pileup count records prior to chunking for fitting background noise distributions.
sequenceContextLength 5 The length of the sequence context bordering the variant on the 5' and 3' ends to be included in the output table.
minimumDepthForHighConfidenceCalls 100 Minimum depth for high-confidence variant calls.
jvmOverhead 192 The memory overhead to allow for the Java Virtual Machine in addition to the memory specified for each Java process.

It is possible to set configuration parameters using command line arguments. This can become unwieldy when changing a large number of settings and the alternative use of a configuration file is generally preferred.

As an example, the path to the reference genome sequence FASTA file, to which the sequence data were aligned, can be specified using the referenceGenomeFasta paramter as follows:

nextflow run crukci-bioinformatics/ampliconseq --referenceGenomeFasta /data/reference_data/reference_genomes/homo_sapiens/GRCh37/fasta/GRCh37.fa

The following example specifies the sample sheet and amplicon coordinates file, and instructs the pipeline to annotate variants using Ensembl VEP for the given species and assembly. It also specifies the variant caller to use (VarDict) and the minimum allele fraction of variants that it can attempt to identify.

nextflow run crukci-bioinformatics/ampliconseq \
    --samples samples.txt \
    --amplicons /data/reference_data/ampliconseq/tp53_panel/amplicons.csv \
    --referenceGenomeFasta /data/reference_data/reference_genomes/homo_sapiens/GRCh37/fasta/GRCh37.fa \
    --vepAnnotation \
    --vepCacheDir /data/reference_data/vep_cache \
    --vepSpecies homo_sapiens \
    --vepAssembly GRCh37 \
    --variantCaller vardict \
    --minimumAlleleFraction 0.01

The default parameter settings can be found in the nextflow.config file that is installed as part of the pipeline.

A more convenient way of setting pipeline parameters makes use of a configuration file, e.g. ampliconseq.config. This is specified when running the pipeline using the -config (or -c) option.

nextflow run crukci-bioinformatics/ampliconseq -c ampliconseq.config

The following is a sample configuration file that sets the same sample sheet, amplicon coordinates file, VEP annotation and variant calling settings as in the above example using command line arguments.

params {
    samples               = "samples.txt"
    amplicons             = "/data/reference_data/ampliconseq/tp53_panel/amplicons.csv"
    referenceGenomeFasta  = "/data/reference_data/reference_genomes/homo_sapiens/GRCh37/fasta/GRCh37.fa"
    vepAnnotation         = true
    vepCacheDir           = "/data/reference_data/vep_cache"
    vepSpecies            = "homo_sapiens"
    vepAssembly           = "GRCh37"
    outputDir             = "results"
    variantCaller         = "vardict"
    minimumAlleleFraction = 0.01
}

This takes the form of a name = value syntax with a separate line for each parameter within curly braces bounding a params block. Note that file names and paths and other character or string values need to be in quotation marks while numeric values do not, and boolean parameters such as vepAnnotation can be set to true or false.

See the Nextflow documentation for more details about the configuration syntax.

The configuration file will normally contain a subset of the parameters specified in the nextflow.config found at the top level of the GitHub repository. nextflow.config contains the default settings, some or all of which are overridden by the configuration file specified with the -config option when running the pipeline.

The pipeline requires a reference genome FASTA file and, optionally, an annotation database or cache for Ensembl Variant Effect Predictor (VEP).

The primary inputs to the pipeline are BAM files containing alignments for sequence reads mapped to a reference genome. The reference genome sequence FASTA file used in the alignment process must be specified using the referenceGenomeFasta parameter. This FASTA file needs to be indexed, e.g. using samtools faidx, and have an accompanying sequence dictionary that can be created with samtools dict or the GATK/Picard CreateSequenceDictionary tool.

The pipeline can annotate variants using Ensembl VEP. It runs VEP in an offline mode using a pre-downloaded annotation cache. The cache for a particular species and genome assembly can be downloaded using a single step supplementary workflow (download_vep_cache) as shown below:

nextflow run crukci-bioinformatics/ampliconseq \
    -main-script download_vep_cache.nf \
    -with-singularity \
    --vepCacheDir /path_to/vep_cache \
    --vepSpecies homo_sapiens \
    --vepAssembly GRCh37

The -with-singularity argument indicates that VEP will be run in a container using Singularity. Use -with-docker to use Docker instead of Singularity or remove the -with-singularity argument if not using a container, in which case the vep_install tool that is installed with VEP will need to be available on the PATH.

Substitute the top-level VEP cache directory as required; note that the download will fail if the directory doesn't already exist.

The VEP cache can be quite large (around 15G for homo sapiens) and downloading and unpacking the cache may take several minutes.

The most straightforward way to run the ampliconseq pipeline is to use the pre-packaged container with either Docker or Singularity by specifying the -with-docker or -with-singularity flag.

nextflow run crukci-bioinformatics/ampliconseq -config ampliconseq.config -with-docker

nextflow run crukci-bioinformatics/ampliconseq -config ampliconseq.config -with-singularity

Nextflow will automatically fetch the container from Docker Hub and will build the Singularity image from the Docker container when running with Singularity. Singularity is more likely than Docker to be available on high-performance cluster computing platforms.

When using Singularity, the pipeline assumes that the user bind control feature is enabled and sets singularity.autoMounts = true in the Nextflow configuration file. See the Nextflow documentation for more details on this.

Alternatively, the use of the Docker container can be specified in the configuration file by adding the following line:

docker.enabled = true

Similarly, to enable Singularity, instead add the following:

singularity.enabled = true

These can also be added as part of an execution profile (see next section).

Resource settings are configured using Nextflow profiles. The ampliconseq pipeline provides three profiles - standard, bigserver and cluster configured for running on servers and the high-performance compute cluster at CRUK CI. These specify the maximum number of CPUs or memory that can be used at any one time during the pipeline run or the maximum number of jobs that can be submitted to the cluster to be run in parallel.

A custom profile can be created in the configuration file, e.g. ampliconseq.config, an example of which is shown below.

profiles {
    myprofile {
        process.executor = 'local'
        executor {
            cpus = 8
            memory = 32.GB
        }
        singularity.enabled = true
    }
}

The new profile, myserver, allows for up to 8 CPUs to be used at any one time and a total of 32G of memory. The pipeline specifies how many CPU cores and how much memory each process requires and Nextflow ensures that the overall resource allocation does not exceed that specified in the profile.

The local executor (the default) runs processes on the computer on which Nextflow is launched. The processes are parallelized by spawning multiple threads and taking advantage of the multi-core architecture provided by the CPU.

Setting singularity.enabled = true in the profile tells Nextflow to use the container with Singularity; it is not necessary to specify this separately with the -with-singularity option.

This profile can be selected by using the -profile command line option.

nextflow run crukci-bioinformatics/ampliconseq -config ampliconseq.config -profile myprofile

The following profile tells Nextflow to submit jobs to nodes on a compute cluster using the Slurm resource manager. It allows for a maximum of 25 jobs to be submitted to the 'long' queue for running in parallel and tells Nextflow to poll every 30 seconds to check for completed jobs.

profiles {
    mycluster {
        process {
            executor = 'slurm'
            queue = 'long'
        }
        executor {
            queueSize = 25
            pollInterval = 30.sec
            jobName = { "'$task.name'" }
        }
        singularity.enabled = true
    }
}

Nextflow can provide a useful summary report detailing the completion status, execution time and memory used by each task, and a timeline chart.

Use the -with-report and -with-timeline command line options to produce these reports, e.g.

 nextflow run crukci-bioinformatics/ampliconseq \
    -config ampliconseq.config \
    -with-report ampliconseq.report.html \
    -with-timeline ampliconseq.timeline.html

Nextflow logs information to a hidden file named .nextflow.log in the launch directory in which ampliconseq is run. This contains logging information that can help with debugging problems with a pipeline run. It will, for example, show which task(s) failed and the directory in which that task was run. An alternative log file name can be specified using the -log command line argument.

nextflow -log ampliconseq.log run crukci-bioinformatics/ampliconseq -config ampliconseq.config

nextflow help gives more details on command line options for Nextflow.

Nextflow runs each task within its own directory. These directories are created under a work directory, by default a subdirectory of the launch directory named work but which is configurable with the -work-dir command line option. Each task directory contains hidden files with names such as .command.sh and .command.out, inspection of which can be helpful when debugging pipeline runs.

Intermediate files created during a pipeline execution are written to the work directories. The final outputs are written either to the launch directory or the directory specified using the --outputDir command line option or the outputDir parameter. The work directory (and all its subdirectories) can be deleted on successful completion of the pipeline unless other Nextflow pipeline runs are also making use of the same top-level work directory.

About

Variant calling pipeline for amplicon sequencing data

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages