Sarek is a workflow designed to detect germline and somatic variants on whole genome, whole exome, or targeted sequencing data.

Initially designed for human and mouse, it can work on any species if a reference genome is available. Sarek is designed to handle single samples, such as single-normal or single-tumor samples, and tumor-normal pairs including additional relapses.

Running the pipeline


The typical command for running the pipeline is as follows:

nextflow run nf-core/sarek --input ./samplesheet.csv --outdir ./results --genome GATK.GRCh38 --tools <TOOLS> -profile docker

This will launch the pipeline and perform variant calling with the tools specified in --tools, see the parameter section for details on variant calling tools. In the above example the pipeline runs with the docker configuration profile. See below for more information about profiles.

Note that the pipeline will create the following files in your working directory:

work                # Directory containing the nextflow working files
<OUTDIR>            # Finished results in specified location (defined with --outdir)
.nextflow_log       # Log file from Nextflow
# Other nextflow hidden files, eg. history of pipeline runs and old logs.

If you wish to repeatedly use the same parameters for multiple runs, rather than specifying each flag in the command, you can specify these in a params file.

Pipeline settings can be provided in a yaml or json file via -params-file <file>.


Do not use -c <file> to specify parameters as this will result in errors. Custom config files specified with -c must only be used for tuning process resource specifications, other infrastructural tweaks (such as output directories), or module arguments (args).

The above pipeline run specified with a params file in yaml format:

nextflow run nf-core/sarek -profile docker -params-file params.yaml

with params.yaml containing:

input: './samplesheet.csv'
outdir: './results/'
genome: 'GRCh37'

You can also generate such YAML/JSON files via nf-core/launch.

Input: Sample sheet configurations

You will need to create a samplesheet with information about the samples you would like to analyse before running the pipeline. Use the parameter --input to specify its location. It has to be a comma-separated file with at least 3 columns, and a header row as shown in the examples below.

It is recommended to use the absolute path of the files, but a relative path should also work.

If necessary, a tumor sample can be associated to a normal sample as a pair, if specified with the same patient ID, a different sample, and the respective status. An additional tumor sample (such as a relapse for example), can be added if specified with the same patient ID, a different sample, and the status value 1.

Sarek will output results in a different directory for each sample. If multiple samples IDs are specified in the CSV file, Sarek will consider all files to be from different samples.

Output from Variant Calling and/or Annotation will be in a specific directory for each sample and tool configuration (or normal/tumor pair if applicable).

Overview: Samplesheet Columns

patientCustom patient ID; designates the patient/subject; must be unique for each patient, but one patient can have multiple samples (e.g. normal and tumor).
sexSex chromosomes of the patient; i.e. XX, XY…, only used for Copy-Number Variation analysis in a tumor/pair
Optional, Default: NA
statusNormal/tumor status of sample; can be 0 (normal) or 1 (tumor).
Optional, Default: 0
sampleCustom sample ID for each tumor and normal sample; more than one tumor sample for each subject is possible, i.e. a tumor and a relapse; samples can have multiple lanes for which the same ID must be used to merge them later (see also lane). Sample IDs must be unique for unique biological samples
laneLane ID, used when the sample is multiplexed on several lanes. Must be unique for each lane in the same sample (but does not need to be the original lane name), and must contain at least one character
Required for --step mapping
fastq_1Full path to FastQ file for Illumina short reads 1. File has to be gzipped and have the extension .fastq.gz or .fq.gz.
fastq_2Full path to FastQ file for Illumina short reads 2. File has to be gzipped and have the extension .fastq.gz or .fq.gz.
bamFull path to (u)BAM file
baiFull path to BAM index file
cramFull path to CRAM file
craiFull path to CRAM index file
tableFull path to recalibration table file
vcfFull path to vcf file

An example samplesheet has been provided with the pipeline.

Start with mapping (--step mapping [default])

This step can be started either from FastQ files or (u)BAMs. The CSV must contain at least the columns patient, sample, lane, and either fastq_1/fastq_2 or bam.


Minimal config file:


In this example, the sample is multiplexed over three lanes:


Full samplesheet

In this example, all possible columns are used. There are three lanes for the normal sample, two for the tumor sample, and one for the relapse sample, including the sex and status information per patient:


Start with duplicate marking (--step markduplicates)

Duplicate Marking

For starting from duplicate marking, the CSV file must contain at least the columns patient, sample, bam, bai or patient, sample, cram, crai

NB: When using GATK4 MarkduplicatesSpark reads should be name-sorted for efficient execution



The Sarek-generated CSV file is stored under results/csv/mapped.csv if in a previous run --save_mapped was set and will automatically be used as an input when specifying the parameter --step markduplicates. Otherwise this file will need to be manually generated.

Full samplesheet

In this example, all possible columns are used including the sex and status information per patient:


Start with preparing the recalibration tables (--step prepare_recalibration)

For starting directly from preparing the recalibration tables, the CSV file must contain at least the columns patient, sample, bam, bai or patient, sample, cram, crai.



The Sarek-generated CSV file is stored under results/csv/markduplicates_no_table.csv and will automatically be used as an input when specifying the parameter --step prepare_recalibration.

Full samplesheet

In this example, all possible columns are used including the sex and status information per patient:


Start with base quality score recalibration (--step recalibrate)

For starting from base quality score recalibration the CSV file must contain at least the columns patient, sample, bam, bai, table or patient, sample, cram, crai, table containing the paths to non-recalibrated CRAM/BAM files and the associated recalibration table.



The Sarek-generated CSV file is stored under results/csv/markduplicates.csv and will automatically be used as an input when specifying the parameter --step recalibrate.

Full samplesheet

In this example, all possible columns are used including the sex and status information per patient:


Start with variant calling (--step variant_calling)

For starting from the variant calling step, the CSV file must contain at least the columns patient, sample, bam, bai or patient, sample, cram, crai.



The Sarek-generated CSV file is stored under results/csv/recalibrated.csv and will automatically be used as an input when specifying the parameter --step variant_calling.

Full samplesheet

In this example, all possible columns are used including the sex and status information per patient:


Start with annotation (--step annotate)

For starting from the annotation step, the CSV file must contain at least the columns patient, sample, vcf.

As Sarek will use bgzip and tabix to compress and index the annotated VCF files, it expects the input VCF files to be sorted and compressed.



The Sarek-generated CSV file is stored under results/csv/variantcalled.csv and will automatically be used as an input when specifying the parameter --step annotation.

Full samplesheet

In this example, all possible columns are used including the variantcaller information per sample:


Updating the pipeline

When you launch a pipeline from the command-line with nextflow run nf-core/sarek -profile docker -params-file params.yaml, Nextflow will automatically pull the pipeline code from GitHub and store it as a cached version. When running the pipeline after this, it will always use the cached version if available - even if the pipeline has been updated since. To make sure that you’re running the latest version of the pipeline, make sure that you regularly update the cached version of the pipeline:

nextflow pull nf-core/sarek


It is a good idea to specify a pipeline version when running the pipeline on your data. This ensures that a specific version of the pipeline code and software are used when you run your pipeline. If you keep using the same tag, you’ll be running the same version of the pipeline, even if there have been changes to the code since.

First, go to the nf-core/sarek releases page and find the latest version number - numeric only (eg. 3.1.1). Then specify this when running the pipeline with -r (one hyphen) - eg. -r 3.1.1. Of course, you can switch to another version by changing the number after the -r flag.

This version number will be logged in reports when you run the pipeline, so that you’ll know what you used when you look back in the future. For example, at the bottom of the MultiQC reports.

To further assist in reproducbility, you can use share and re-use parameter files to repeat pipeline runs with the same settings without having to write out a command with every single parameter.


If you wish to share such profile (such as upload as supplementary material for academic publications), make sure to NOT include cluster specific paths to files, nor institutional specific profiles.

Core Nextflow arguments


These options are part of Nextflow and use a single hyphen (pipeline parameters use a double-hyphen).


Use this parameter to choose a configuration profile. Profiles can give configuration presets for different compute environments.

Several generic profiles are bundled with the pipeline which instruct the pipeline to use software packaged using different methods (Docker, Singularity, Podman, Shifter, Charliecloud, Apptainer, Conda) - see below.


We highly recommend the use of Docker or Singularity containers for full pipeline reproducibility, however when this is not possible, Conda is also supported.

The pipeline also dynamically loads configurations from when it runs, making multiple config profiles for various institutional clusters available at run time. For more information and to see if your system is available in these configs please see the nf-core/configs documentation.

Note that multiple profiles can be loaded, for example: -profile test,docker - the order of arguments is important! They are loaded in sequence, so later profiles can overwrite earlier profiles.

If -profile is not specified, the pipeline will run locally and expect all software to be installed and available on the PATH. This is not recommended, since it can lead to different results on different machines dependent on the computer enviroment.

  • test
    • A profile with a complete configuration for automated testing
    • Includes links to test data so needs no other parameters
  • docker
    • A generic configuration profile to be used with Docker
  • singularity
    • A generic configuration profile to be used with Singularity
  • podman
    • A generic configuration profile to be used with Podman
  • shifter
    • A generic configuration profile to be used with Shifter
  • charliecloud
    • A generic configuration profile to be used with Charliecloud
  • apptainer
    • A generic configuration profile to be used with Apptainer
  • conda
    • A generic configuration profile to be used with Conda. Please only use Conda as a last resort i.e. when it’s not possible to run the pipeline with Docker, Singularity, Podman, Shifter, Charliecloud, or Apptainer.


Specify this when restarting a pipeline. Nextflow will use cached results from any pipeline steps where the inputs are the same, continuing from where it got to previously. For input to be considered the same, not only the names must be identical but the files’ contents as well. For more info about this parameter, see this blog post.

You can also supply a run name to resume a specific run: -resume [run-name]. Use the nextflow log command to show previous run names.


Specify the path to a specific config file (this is a core Nextflow command). See the nf-core website documentation for more information.

Nextflow memory requirements

In some cases, the Nextflow Java virtual machines can start to request a large amount of memory. We recommend adding the following line to your environment to limit this (typically in ~/.bashrc or ~./bash_profile):

NXF_OPTS='-Xms1g -Xmx4g'

Running in the background

Nextflow handles job submissions and supervises the running jobs. The Nextflow process must run until the pipeline is finished.

The Nextflow -bg flag launches Nextflow in the background, detached from your terminal so that the workflow does not stop if you log out of your session. The logs are saved to a file.

Alternatively, you can use screen / tmux or similar tool to create a detached session which you can log back into at a later time. Some HPC setups also allow you to run nextflow within a cluster job submitted your job scheduler (from where it submits more jobs).

Custom configuration

Resource requests

Whilst the default requirements set within the pipeline will hopefully work for most people and with most input data, you may find that you want to customise the compute resources that the pipeline requests. Each step in the pipeline has a default set of requirements for number of CPUs, memory and time. For most of the steps in the pipeline, if the job exits with any of the error codes specified here it will automatically be resubmitted with higher requests (2 x original, then 3 x original). If it still fails after the third attempt then the pipeline execution is stopped.

To change the resource requests, please see the max resources and tuning workflow resources section of the nf-core website.

Custom Containers

In some cases you may wish to change which container or conda environment a step of the pipeline uses for a particular tool. By default nf-core pipelines use containers and software from the biocontainers or bioconda projects. However in some cases the pipeline specified version maybe out of date.

To use a different container from the default container or conda environment specified in a pipeline, please see the updating tool versions section of the nf-core website.

Custom Tool Arguments

A pipeline might not always support every possible argument or option of a particular tool used in pipeline. Fortunately, nf-core pipelines provide some freedom to users to insert additional parameters that the pipeline does not include by default.

To learn how to provide additional arguments to a particular tool of the pipeline, please see the customising tool arguments section of the nf-core website.


In most cases, you will only need to create a custom config as a one-off but if you and others within your organisation are likely to be running nf-core pipelines regularly and need to use the same settings regularly it may be a good idea to request that your custom config file is uploaded to the nf-core/configs git repository. Before you do this please can you test that the config file works with your pipeline of choice using the -c parameter. You can then create a pull request to the nf-core/configs repository with the addition of your config file, associated documentation file (see examples in nf-core/configs/docs), and amending nfcore_custom.config to include your custom profile.

See the main Nextflow documentation for more information about creating your own configuration files.

If you have any questions or issues please send us a message on Slack on the #configs channel.

Azure Resource Requests

To be used with the azurebatch profile by specifying the -profile azurebatch. We recommend providing a compute params.vm_type of Standard_D16_v3 VMs by default but these options can be changed if required.

Note that the choice of VM size depends on your quota and the overall workload during the analysis. For a thorough list, please refer the Azure Sizes for virtual machines in Azure.

Troubleshooting & FAQ

How to test the pipeline

When using default parameters only, sarek runs preprocessing and Strelka2. This is reflected in the default test profile:

nextflow run nf-core/sarek -r 3.2.1 -profile test,<container/institute> --outdir results

Expected run output:

[85/6b7739] process > NFCORE_SAREK:SAREK:PREPARE_GENOME:BWAMEM1_INDEX (genome.fasta)                                                [100%] 1 of 1 ✔
[-        ] process > NFCORE_SAREK:SAREK:PREPARE_GENOME:BWAMEM2_INDEX                                                               -
[-        ] process > NFCORE_SAREK:SAREK:PREPARE_GENOME:DRAGMAP_HASHTABLE                                                           -
[22/cf54a8] process > NFCORE_SAREK:SAREK:PREPARE_GENOME:GATK4_CREATESEQUENCEDICTIONARY (genome.fasta)                               [100%] 1 of 1 ✔
[-        ] process > NFCORE_SAREK:SAREK:PREPARE_GENOME:MSISENSORPRO_SCAN                                                           -
[28/dad25a] process > NFCORE_SAREK:SAREK:PREPARE_GENOME:SAMTOOLS_FAIDX (genome.fasta)                                               [100%] 1 of 1 ✔
[23/3fe964] process > NFCORE_SAREK:SAREK:PREPARE_GENOME:TABIX_DBSNP (dbsnp_146.hg38.vcf)                                            [100%] 1 of 1 ✔
[-        ] process > NFCORE_SAREK:SAREK:PREPARE_GENOME:TABIX_GERMLINE_RESOURCE                                                     -
[-        ] process > NFCORE_SAREK:SAREK:PREPARE_GENOME:TABIX_KNOWN_SNPS                                                            -
[14/26e286] process > NFCORE_SAREK:SAREK:PREPARE_GENOME:TABIX_KNOWN_INDELS (mills_and_1000G.indels.vcf)                             [100%] 1 of 1 ✔
[-        ] process > NFCORE_SAREK:SAREK:PREPARE_GENOME:TABIX_PON                                                                   -
[76/04d107] process > NFCORE_SAREK:SAREK:PREPARE_INTERVALS:CREATE_INTERVALS_BED (genome.interval_list)                              [100%] 1 of 1 ✔
[d4/f97174] process > NFCORE_SAREK:SAREK:PREPARE_INTERVALS:GATK4_INTERVALLISTTOBED (genome)                                         [100%] 1 of 1 ✔
[70/82ba3c] process > NFCORE_SAREK:SAREK:PREPARE_INTERVALS:TABIX_BGZIPTABIX_INTERVAL_SPLIT (chr22_1-40001)                          [100%] 1 of 1 ✔
[d4/c2d0c4] process > NFCORE_SAREK:SAREK:PREPARE_INTERVALS:TABIX_BGZIPTABIX_INTERVAL_COMBINED (genome)                              [100%] 1 of 1 ✔
[-        ] process > NFCORE_SAREK:SAREK:CONVERT_FASTQ_INPUT:SAMTOOLS_VIEW_MAP_MAP                                                  -
[-        ] process > NFCORE_SAREK:SAREK:CONVERT_FASTQ_INPUT:SAMTOOLS_VIEW_UNMAP_UNMAP                                              -
[-        ] process > NFCORE_SAREK:SAREK:CONVERT_FASTQ_INPUT:SAMTOOLS_VIEW_UNMAP_MAP                                                -
[-        ] process > NFCORE_SAREK:SAREK:CONVERT_FASTQ_INPUT:SAMTOOLS_VIEW_MAP_UNMAP                                                -
[-        ] process > NFCORE_SAREK:SAREK:CONVERT_FASTQ_INPUT:SAMTOOLS_MERGE_UNMAP                                                   -
[-        ] process > NFCORE_SAREK:SAREK:CONVERT_FASTQ_INPUT:COLLATE_FASTQ_UNMAP                                                    -
[-        ] process > NFCORE_SAREK:SAREK:CONVERT_FASTQ_INPUT:COLLATE_FASTQ_MAP                                                      -
[-        ] process > NFCORE_SAREK:SAREK:CONVERT_FASTQ_INPUT:CAT_FASTQ                                                              -
[c4/f59e5a] process > NFCORE_SAREK:SAREK:FASTQC (test-test_L1)                                                                      [100%] 1 of 1 ✔
[0b/c5a999] process > NFCORE_SAREK:SAREK:FASTQ_ALIGN_BWAMEM_MEM2_DRAGMAP:BWAMEM1_MEM (test)                                         [100%] 1 of 1 ✔
[-        ] process > NFCORE_SAREK:SAREK:FASTQ_ALIGN_BWAMEM_MEM2_DRAGMAP:BWAMEM2_MEM                                                -
[-        ] process > NFCORE_SAREK:SAREK:FASTQ_ALIGN_BWAMEM_MEM2_DRAGMAP:DRAGMAP_ALIGN                                              -
[c7/664cd1] process > NFCORE_SAREK:SAREK:BAM_MARKDUPLICATES:GATK4_MARKDUPLICATES (test)                                             [100%] 1 of 1 ✔
[13/bc73b6] process > NFCORE_SAREK:SAREK:BAM_MARKDUPLICATES:INDEX_MARKDUPLICATES (test)                                             [100%] 1 of 1 ✔
[2a/99608e] process > NFCORE_SAREK:SAREK:BAM_MARKDUPLICATES:CRAM_QC_MOSDEPTH_SAMTOOLS:SAMTOOLS_STATS (test)                         [100%] 1 of 1 ✔
[f2/0420ca] process > NFCORE_SAREK:SAREK:BAM_MARKDUPLICATES:CRAM_QC_MOSDEPTH_SAMTOOLS:MOSDEPTH (test)                               [100%] 1 of 1 ✔
[-        ] process > NFCORE_SAREK:SAREK:CRAM_TO_BAM                                                                                -
[eb/46945a] process > NFCORE_SAREK:SAREK:BAM_BASERECALIBRATOR:GATK4_BASERECALIBRATOR (test)                                         [100%] 1 of 1 ✔
[-        ] process > NFCORE_SAREK:SAREK:BAM_BASERECALIBRATOR:GATK4_GATHERBQSRREPORTS                                               -
[ec/2377d4] process > NFCORE_SAREK:SAREK:BAM_APPLYBQSR:GATK4_APPLYBQSR (test)                                                       [100%] 1 of 1 ✔
[-        ] process > NFCORE_SAREK:SAREK:BAM_APPLYBQSR:CRAM_MERGE_INDEX_SAMTOOLS:MERGE_CRAM                                         -
[88/3af664] process > NFCORE_SAREK:SAREK:BAM_APPLYBQSR:CRAM_MERGE_INDEX_SAMTOOLS:INDEX_CRAM (test)                                  [100%] 1 of 1 ✔
[f4/828fde] process > NFCORE_SAREK:SAREK:CRAM_QC_RECAL:SAMTOOLS_STATS (test)                                                        [100%] 1 of 1 ✔
[fb/a9d66f] process > NFCORE_SAREK:SAREK:CRAM_QC_RECAL:MOSDEPTH (test)                                                              [100%] 1 of 1 ✔
[-        ] process > NFCORE_SAREK:SAREK:CRAM_TO_BAM_RECAL                                                                          -
[bc/f3f5cf] process > NFCORE_SAREK:SAREK:VCF_QC_BCFTOOLS_VCFTOOLS:BCFTOOLS_STATS (test)                                             [100%] 1 of 1 ✔
[21/8d4f02] process > NFCORE_SAREK:SAREK:VCF_QC_BCFTOOLS_VCFTOOLS:VCFTOOLS_TSTV_COUNT (test)                                        [100%] 1 of 1 ✔
[36/957fba] process > NFCORE_SAREK:SAREK:VCF_QC_BCFTOOLS_VCFTOOLS:VCFTOOLS_TSTV_QUAL (test)                                         [100%] 1 of 1 ✔
[70/a8e064] process > NFCORE_SAREK:SAREK:VCF_QC_BCFTOOLS_VCFTOOLS:VCFTOOLS_SUMMARY (test)                                           [100%] 1 of 1 ✔
[36/e35b1b] process > NFCORE_SAREK:SAREK:CUSTOM_DUMPSOFTWAREVERSIONS (1)                                                            [100%] 1 of 1 ✔
[3f/3c3356] process > NFCORE_SAREK:SAREK:MULTIQC                                                                                    [100%] 1 of 1 ✔
-[nf-core/sarek] Pipeline completed successfully-
Completed at: 09-Jun-2023 13:46:31
Duration    : 1m 50s
CPU hours   : (a few seconds)
Succeeded   : 27

The pipeline comes with a number of possible paths and tools that can be used.

Due to the small test data size, unfortunately not everything can be tested from top-to-bottom, but often is done by utilizing the pipeline’s --step parameter.

For more extensive testing purpose, we have the test_cache profile that contain the same data, but on which the path to the reference and input files can be changed using the --test_data_base params.

Annotation is generally tested separately from the remaining workflow, since we use references for C.elegans, while the remaining tests are run on downsampled human data.

nextflow run nf-core/sarek -r 3.2.1 -profile test_cache,<container/institute> --outdir results --tools snpeff --step annotation

If you are interested in any of the other tests that are run on every code change or would like to run them yourself, you can take a look at tests/<filename>.yml. For each entry the respective nextflow command run and the expected output is specified.

Some of the currently, available test profiles:

Test profileRun command
annotationnextflow run -profile test_cache,annotation,docker --tools snpeff,vep,merge
no_intervalsnextflow run -profile test_cache,no_intervals,docker
targetednextflow run -profile test_cache,targeted,docker
tools_germlinenextflow run -profile test_cache,tools_germline,docker --tools strelka
tools_tumoronlynextflow run -profile test_cache,tools_tumoronly,docker --tools strelka
tools_somaticnextflow run -profile test_cache,tools_somatic,docker --tools strelka
trimmingnextflow run -profile test_cache,trim_fastq,docker
uminextflow run -profile test_cache,umi,docker
use_gatk_sparknextflow run -profile test_cache,use_gatk_spark,docker

If you are interested in any of the other profiles that are used, you can take a look at conf/test/<filename>.config.

How can the different steps be used

Sarek can be started at different points in the analysis by setting the parameter --step. Once started at a certain point, the pipeline runs through all the following steps without additional intervention. For example when starting from --step mapping (set by default) and --tools strelka,vep, the input reads will be aligned, duplicate marked, recalibrated, variant called with Strelka, and finally VEP will annotate the called variants.

Which variant calling tool is implemented for which data type?

This list is by no means exhaustive and it will depend on the specific analysis you would like to run. This is a suggestion based on the individual docs of the tools specifically for human genomes and a garden-variety sequencing run as well as what has been added to the pipeline.

ToolWGSWES Panel NormalTumorSomatic
GATK HaplotypeCallerxxxx--
GATK Mutect2xxx-xx

How to run ASCAT with whole-exome sequencing data?

ASCAT runs out of the box on whole genome sequencing data using iGenomes resources. While the ASCAT implementation in sarek is capable of running with whole-exome sequencing data, the needed references are currently not provided with the igenomes.config. According to the developers of ASCAT, loci and allele files (one file per chromosome) can be downloaded directly from the Battenberg repository.

Please note that:

  • Row names (for GC and RT correction files) should be ${chr}_${position} (there is no SNP/probe ID for HTS data).
  • All row names in GC and RT correction files should also appear in the loci files
  • Loci and allele files must contain the same set of SNPs
  • ASCAT developers strongly recommend using a BED file for WES/TS data. This prevents considering SNPs covered by off-target reads that would add noise to log/BAF tracks.
  • The total number of GC correction loci in a sample must be at least 10% of the number of loci with logR values. If the number of GC correction loci is too small compared to the total number of loci, ASCAT will throw an error.

From ‘Reference files’

For WES and targeted sequencing, we recommend using the reference files (loci, allele and logR correction files) as part of the Battenberg package. Because they require a high-resolution input, our reference files for WGS are not suitable for WES and targeted sequencing. For WES, loci and allele files from the Battenberg package can be fed into ascat.prepareHTS. For targeted sequencing, allele files from the Battenberg package can be fed into ascat.prepareTargetedSeq, which will generate cleaned loci and allele files that can be fed into ascat.prepareHTS.

How to generate ASCAT resources for exome or targeted sequencing

  1. Fetch the GC content correction and replication timing (RT) correction files from the Dropbox links provided by the ASCAT developers and intersect the SNP coordinates with the exome target coordinates. If the target file has ‘chr’ prefixes, make a copy with these removed first. Extract the GC and RT information for only the on target SNPs and zip the results.
sed -e 's/chr//' targets_with_chr.bed > targets.bed
for t in GC RT
  unzip ${t}
  cut -f 1-3 ${t}_G1000_hg38.txt > ascat_${t}_snps_hg38.txt
  tail -n +2 ascat_${t}_snps_hg38.txt | awk '{ print $2 "\t" $3-1 "\t" $3 "\t" $1 }' > ascat_${t}_snps_hg38.bed
  bedtools intersect -a ascat_${t}_snps_hg38.bed -b targets.bed | awk '{ print $1 "_" $3 }' > ascat_${t}_snps_on_target_hg38.txt
  head -n 1 ${t}_G1000_hg38.txt > ${t}_G1000_on_target_hg38.txt
  grep -f ascat_${t}_snps_on_target_hg38.txt ${t}_G1000_hg38.txt >> ${t}_G1000_on_target_hg38.txt
  zip ${t} ${t}_G1000_on_target_hg38.txt
  rm ${t}
  1. Download the Battenberg 1000G loci and alleles files. The steps below follow downloading from the Battenberg repository at the Oxford University Research Archive. The files are also available via Dropbox links from the same page as the GC and RT correction files above.
mv rt345gd52w
tar xf
cd 1000G_loci_hg38
mkdir battenberg_alleles_on_target_hg38
mv *allele* battenberg_alleles_on_target_hg38/
mkdir battenberg_loci_on_target_hg38
mv *loci* battenberg_loci_on_target_hg38/
  1. Copy the targets_with_chr.bed and GC_G1000_on_target_hg38.txt files into the newly created battenberg_loci_on_target_hg38 folder before running the next set of steps. ASCAT generates a list of GC correction loci with sufficient coverage in a sample, then intersects that with the list of all loci with tumour logR values in that sample. If the intersection is <10% the size of the latter, it will fail with an error. Because the Battenberg loci/allele sets are very dense, subsetting to on-target regions is still too many loci. This script ensures that all SNPs with GC correction information are included in the loci list, plus a random sample of another 30% of all on target loci. You may need to vary this proportion depending on your set of targets. A good rule of thumb is that the size of your GC correction loci list should be about 15% the size of your total loci list. This allows for a margin of error.
cd battenberg_loci_on_target_hg38/
rm *chrstring*
rm 1kg.phase3.v5a_GRCh38nounref_loci_chr23.txt
for i in {1..22} X
   awk '{ print $1 "\t" $2-1 "\t" $2 }' 1kg.phase3.v5a_GRCh38nounref_loci_chr${i}.txt > chr${i}.bed
   grep "^${i}_" GC_G1000_on_target_hg38.txt | awk '{ print "chr" $1 }' > chr${i}.txt
   bedtools intersect -a chr${i}.bed -b targets_with_chr.bed | awk '{ print $1 "_" $3 }' > chr${i}_on_target.txt
   n=`wc -l chr${i}_on_target.txt | awk '{ print $1 }'`
   count=$((n * 3 / 10))
   grep -xf chr${i}.txt chr${i}_on_target.txt > chr${i}.temp
   shuf -n $count chr${i}_on_target.txt >> chr${i}.temp
   sort -n -k2 -t '_' chr${i}.temp | uniq | awk 'BEGIN { FS="_" } ; { print $1 "\t" $2 }' > battenberg_loci_on_target_hg38_chr${i}.txt
zip battenberg_loci_on_target_hg38_chr*.txt
  1. Extract the alleles for the same set of SNPs. Uses a short R script defined below.
cd ../battenberg_alleles_on_target_hg38/
rm 1kg.phase3.v5a_GRCh38nounref_allele_index_chr23.txt
for i in {1..22} X
  Rscript intersect_ascat_alleles.R ../battenberg_loci_on_target_hg38/battenberg_loci_on_target_hg38_chr${i}.txt \
    1kg.phase3.v5a_GRCh38nounref_allele_index_chr${i}.txt battenberg_alleles_on_target_hg38_chr${i}.txt
zip battenberg_alleles_on_target_hg38_chr*.txt

Rscript intersect_ascat_alleles.R

#!/usr/bin/env Rscript
args = commandArgs(trailingOnly=TRUE)
loci = read.table(args[1], header=F, sep="\t", stringsAsFactors=F)
alleles = read.table(args[2], header=T, sep="\t", stringsAsFactors=F)
i = intersect(loci$V2, alleles$position)
out = subset(alleles, alleles$position %in% i)
write.table(out, args[3], col.names=T, row.names=F, quote=F, sep="\t")
  1. Move or copy all of the zip files you’ve created to a suitable location. Specify these in your parameters, e.g.
  "ascat_alleles": "/path/to/",
  "ascat_loci": "/path/to/",
  "ascat_loci_gc": "/path/to/",
  "ascat_loci_rt": "/path/to/"

What are the bwa, bwa-mem2 and sentieon bwa mem parameters?

For mapping, sarek follows the parameter suggestions provided in this paper:

-K 100000000 : for deterministic pipeline results, for more info see here

-Y: force soft-clipping rather than default hard-clipping of supplementary alignments

In addition, currently the mismatch penalty for reads with tumor status in the sample sheet are mapped with a mismatch penalty of -B 3.

How to manage scatter/gathering (parallelization with-in each sample)

While Nextflow ensures all samples are run in parallel, the pipeline can split input files for each sample into smaller chunks which are processes in parallel. This speeds up analysis for individual chunks, but might occupy more storage space.

Therefore, the different scatter/gather options can be set by the user:

Split Fastq files

By default, the input fastq files are split into smaller chunks with FASTP, mapped in parallel, and then merged and duplicate marked. This can be customized by setting the parameter --split_fastq. This parameter determines how many reads are within each split. Setting it to 0 will turn of any splitting and only one mapping process is run per input fastq file.

FastP creates as many chunks as CPUs are specified (by default 12) and subdivides them further, if the number of reads in a chunk is larger then the value specified in --split_fastq. Thus, the parameter --split_fastq is an upper bound, e.g. if 1/12th of the Fastq file exceeds the provided value another fastq file will be generated.

Intervals for Base Quality Score Recalibration and Variantcalling

The pipeline can parallelize base quality score recalibration and variant calling across genomic chunks of roughly similar sizes. For this, a bed file containing genomic regions of interest is used, it’s the intervals file. By default, the intervals file for WGS used is the one provided by GATK (details here). When running targeted analysis, it is recommended to use the bed file containing the targeted regions.

The amount of scatter/gathering can be customized by adjusting the parameter --nucleotides_per_second.

NB: The same intervals are processed regardless of the number of groups. The number of groups however determines over how many compute nodes the analysis is scattered on.

The default value is 200000, increasing this value will reduce the number of groups that are processed in parallel. Generally, smaller numbers of groups (each group has more regions), the slower the processing, and less storage space is consumed. In particular, in cloud computing setting it is often advisable to reduce the number of groups to be run in parallel to reduce data staging steps.

How to create a panel-of-normals for Mutect2

For a detailed tutorial on how to create a panel-of-normals, see here.

If you have problems running processes that make use of Spark such as MarkDuplicates. You are probably experiencing issues with the limit of open files in your system. You can check your current limit by typing the following:

ulimit -n

The default limit size is usually 1024 which is quite low to run Spark jobs. In order to increase the size limit permanently you can:

Edit the file /etc/security/limits.conf and add the lines:

*     soft   nofile  65535
*     hard   nofile  65535

Edit the file /etc/sysctl.conf and add the line:

fs.file-max = 65535

Edit the file /etc/sysconfig/docker and add the new limits to OPTIONS like this:

OPTIONS=”—default-ulimit nofile=65535:65535"

Re-start your session.

Note that the way to increase the open file limit in your system may be slightly different or require additional steps.

Cannot delete work folder when using docker + Spark

Currently, when running spark-based tools in combination with docker, it is required to set docker.userEmulation = false. This can unfortunately causes permission issues when work/ is being written with root permissions. In case this happens, you might need to configure docker to run without userEmulation (see here).

How to handle UMIs

Sarek can process UMI-reads, using fgbio tools.

In order to use reads containing UMI tags as your initial input, you need to include --umi_read_structure [structure] in your parameters.

This will enable pre-processing of the reads and UMI consensus reads calling, which will then be used to continue the workflow from the mapping steps. For post-UMI processing depending on the experimental setup, duplicate marking and base quality recalibration can be skipped with [--skip_tools].

UMI Read Structure

This parameter is a string, which follows a convention to describe the structure of the umi.

As an example: if your reads contain a UMI only on the forward read, the string can only represent one structure (i.e. “2M11S+T”); should your reads contain a UMI on both reas, the string will contain two structures separated by a blank space (i.e. “2M11S+T 2M11S+T”); should your reads contain a UMI only on the reverse read, your structure must represent the template only for the forward read and template plus UMI for the reverse read (i.e. +T 12M11S+T). Please do refer to FGBIO documentation for more details, as providing the correct structure is essential and specific to the UMI kit used.

Limitations and future updates

Recent updates to Samtools have been introduced, which can speed-up performance of fgbio tools used in this workflow. The current workflow does not handle duplex UMIs (i.e. where opposite strands of a duplex molecule have been tagged with a different UMI), and best practices have been proposed to process this type of data. Both changes will be implemented in a future release.

How to run sarek when no(t all) reference files are in igenomes

For common genomes, such as GRCh38 and GRCh37, the pipeline is shipped with (almost) all necessary reference files. However, sometimes it is necessary to use custom references for some or all files:

No igenomes reference files are used

If none of your required genome files are in igenomes, --igenomes_ignore must be set to ignore any igenomes input and --genome null. The fasta file is the only required input file and must be provided to run the pipeline. All other possible reference file can be provided in addition. For details, see the paramter documentation.

Minimal example for custom genomes:

nextflow run nf-core/sarek --genome null --igenomes_ignore --fasta <custom.fasta>

Overwrite specific reference files

If you don’t want to use some of the provided reference genomes, they can be overwritten by either providing a new file or setting the respective file parameter to false, if it should be ignored:

Example for using a custom known indels file:

nextflow run nf-core/sarek --known_indels <my_known_indels.vcf.gz> --genome GRCh38.GATK

Example for not using known indels, but all other provided reference file:

nextflow run nf-core/sarek --known_indels false --genome GRCh38.GATK

Where do the used reference genomes originate from

For GATK.GRCh38 the links for each reference file and the corresponding processes that use them is listed below. For GATK.GRCh37 the files originate from the same sources:

bwabwa-membwa index -p bwa/fasta.baseName{fasta.baseName} fasta
bwamem2bwa-mem2bwa-mem2 index -p bwamem2/fasta{fasta} fasta
dragmapDragMapdragen-os —build-hash-table true —ht-reference $fasta —output-directory dragmap
dbsnpBaserecalibrator, ControlFREEC, GenotypeGVCF, HaplotypeCallerGATKBundle
dbsnp_tbiBaserecalibrator, ControlFREEC, GenotypeGVCF, HaplotypeCallerGATKBundle
dictBaserecalibrator(Spark), CNNScoreVariant, EstimateLibraryComplexity, FilterMutectCalls, FilterVariantTranches, GatherPileupSummaries,GenotypeGVCF, GetPileupSummaries, HaplotypeCaller, MarkDulpicates(Spark), MergeVCFs, Mutect2, VariantrecalibratorGATKBundle
fastaApplyBQSR(Spark), ApplyVQSR, ASCAT, Baserecalibrator(Spark), BWA, BWAMem2, CNNScoreVariant, CNVKit, ControlFREEC, DragMap, DEEPVariant, EnsemblVEP, EstimateLibraryComplexity, FilterMutectCalls, FilterVariantTranches, FreeBayes, GatherPileupSummaries,GenotypeGVCF, GetPileupSummaries, HaplotypeCaller, interval building, Manta, MarkDuplicates(Spark),MergeVCFs,MSISensorPro, Mutect2, Samtools, SnpEff, Strelka, Tiddit, VariantrecalibratorGATKBundle
fasta_faiApplyBQSR(Spark), ApplyVQSR, ASCAT, Baserecalibrator(Spark), BWA, BWAMem2, CNNScoreVariant, CNVKit, ControlFREEC, DragMap, DEEPVariant, EnsemblVEP, EstimateLibraryComplexity, FilterMutectCalls, FilterVariantTranches, FreeBayes, GatherPileupSummaries,GenotypeGVCF, GetPileupSummaries, HaplotypeCaller, interval building, Manta, MarkDuplicates(Spark),MergeVCFs,MSISensorPro, Mutect2, Samtools, SnpEff, Strelka, Tiddit, VariantrecalibratorGATKBundle
intervalsApplyBQSR(Spark), ASCAT, Baserecalibrator(Spark), BCFTools, CNNScoreVariants, ControlFREEC, Deepvariant, FilterVariantTranches, FreeBayes, GenotypeGVCF, GetPileupSummaries, HaplotypeCaller, Strelka, mpileup, MSISensorPro, Mutect2, VCFToolsGATKBundle
known_indelsBaseRecalibrator(Spark), FilterVariantTranchesGATKBundle
known_indels_tbiBaseRecalibrator(Spark), FilterVariantTranchesGATKBundle
known_snpsBaseRecalibrator(Spark), FilterVariantTranches, VariantRecalibratorGATKBundle
known_snps_tbiBaseRecalibrator(Spark), FilterVariantTranches, VariantRecalibratorGATKBundle

How to customise SnpEff and VEP annotation

SNPeff and VEP both require a large resource of files known as a cache. These are folders composed of multiple gigabytes of files which need to be available for the software to properly function. To use these, supply the parameters --vep_cache and/or --snpeff_cache with the locations to the root of the annotation cache folder for each tool.

Specify the cache location

Params --snpeff_cache and --vep_cache are used to specify the locations to the root of the annotation cache folder. The cache will be located within a subfolder with the path ${snpeff_species}.${snpeff_version} for SnpEff and ${vep_species}/${vep_genome}_${vep_cache_version} for VEP. If this directory is missing, Sarek will raise an error.

For example this is a typical folder structure for GRCh38 and WBCel235, with SNPeff cache version 105 and VEP cache version 110:

├─ snpeff_cache/
│  ├─ GRCh38.105/
│  ├─ WBcel235.105/
├─ vep_cache/
│  ├─ caenorhabditis_elegans/
│  │  ├─ 110_WBCel235/
│  ├─ homo_sapiens/
│  │  ├─ 110_GRCh38/

For this example, the parameters --snpeff_cache /data/snpeff_cache and --vep_cache /data/vep_cache would be used. Both SnpEff and VEP will figure out internally the path towards the specific cache version / species the annotation should be performed given the parameters specified to Sarek.

Change cache version and species

By default all is specified in the igenomes.config file. Explanation can be found for all params in the documentation:

With the previous example of GRCh38, these are the values that were used for these params:

snpeff_db         = '105'
snpeff_genome     = 'GRCh38'
vep_cache_version = '110'
vep_genome        = 'GRCh38'
vep_species       = 'homo_sapiens'

Usage recommendation with AWS iGenomes

The cache for each of these annotation tools has its own structure and is frequently updated, therefore it is kept separate from AWS iGenomes. It is not recommended to put any cache for each of this annotation tools in your local AWS iGenomes folder.

A classical organisation on a shared storage area might be:


Which can then be used this way in Sarek:

nextflow run nf-core/sarek \
    --igenomes_base /data/igenomes/ \
    --snpeff_cache /data/cache/snpeff_cache/ \
    --vep_cache /data/cache/vep_cache/ \

Alternatively the data may be stored on AWS S3 storage, therefore the parameters might be:


Which can then be used this way in Sarek:

nextflow run nf-core/sarek \
    --igenomes_base s3://my-reference-data/igenomes/ \
    --snpeff_cache s3://my-reference-data/cache/ensemblvep/ \
    --vep_cache s3://my-reference-data/cache/snpeff/ \

These params can be specified in a config file or in a profile using the params scope, or even in a json or a yaml file using the -params-file nextflow option.

Note: we recommend storing each annotation cache in a separate directory so each cache version is handled differently. This may mean you have many similar directories but will dramatically reduce the storage burden on machines running the SnpEff or VEP process.

Use annotation-cache for SnpEff and VEP

Annotation-cache is an open AWS registry resource that stores a mirror of some cache files on AWS S3 which can be used with Sarek. It contains some genome builds which can be found by checking the contents of the S3 bucket.

SNPeff and VEP cache are stored at the following location on S3:

snpeff_cache = s3://annotation-cache/snpeff_cache/
vep_cache = s3://annotation-cache/vep_cache/

The contents of said cache can be listed with the following command using the S3 CLI:

aws s3 --no-sign-request ls s3://annotation-cache/snpeff_cache
aws s3 --no-sign-request ls s3://annotation-cache/vep_cache/

Since both Snpeff and VEP are internally figuring the path towards the specific cache version / species, annotation-cache is using an extra set of keys to specify the species and genome build.

So if you are using this resource, please either set --use_annotation_cache_keys to use the AWS annotation cache, or point towards your own cache folder structure matching the expected structure.

Please refer to the annotation-cache documentation for more details.

Use Sarek to download cache and annotate in one go

Both VEP and snpEff come with built-in download functionality to download the cache prior to use. Sarek includes these as optional processes. Use the params --download_cache, and specify the tool with --tools and Sarek will download the relevant cache (snpeff and/or vep) using their respective download functions. It is recommended to save the cache somewhere highly accessible for subsequent runs of Sarek, so the cache does not have to be re-downloaded.

Sarek will automatically download the cache using each tools (SnpEff and/or VEP) to your work directory. And subsequently perform the annotation of VCF files specified as an input in a samplesheet or produced by Sarek.

Only download cache

Using the params --build_only_index allow for only downloading the cache for the specified tools.

Location for the cache

Cache can be downloaded in the specified --outdir_cache location. Else, it will be downloaded in cache/ in the specified --outdir location.

This command could be used to download the cache for both tools in the specified --outdir_cache location:

nextflow run nf-core/sarek -r 3.3.0 --outdir results --outdir_cache /path_to/my-own-cache --tools vep,snpeff --download_cache --build_only_index --input false

This command could be used to point to the recently downloaded cache and run SnpEff and VEP:

nextflow run nf-core/sarek -r 3.3.0 --outdir results --vep_cache /path_to/my-own-cache/vep_cache --snpeff_cache /path_to/my-own-cache/snpeff_cache --tools vep,snpeff --input samplesheet_vcf.csv

Create containers with pre-downloaded cache

nf-core is no longer maintaining containers with pre-downloaded cache. Hosting the cache within the container is not recommended as it can cause a number of problems. Instead we recommned using an external cache. The following is left for legacy reasons.

But for each of these tools, an helper script can be found at the root of the tool folder in the nf-core module repo (snpeff and ensemblvep), and can be adapted for your usage.

Overwritting the container declaration is then possible to accomodate for the new container.

Using VEP plugins


Enable with --vep_dbnsfp. The following parameters are mandatory:

  • --dbnsfp, to specify the path to the dbNSFP processed file.
  • --dbnsfp_tbi, to specify the path to the dbNSFP tabix indexed file.

The following parameters are optionnal:

  • --dbnsfp_consequence, to filter/limit outputs to a specific effect of the variant.
    • The set of consequence terms is defined by the Sequence Ontology and an overview of those used in VEP can be found here.
    • If one wants to filter using several consequences, then separate those by using ’&’ (i.e. --dbnsfp_consequence '3_prime_UTR_variant&intron_variant'.”,
  • --dbnsfp_fields, to retrieve individual values from the dbNSFP file.
    • The values correspond to the name of the columns in the dbNSFP file and are separated by comma.
    • The column names might differ between the different dbNSFP versions. Please check the Readme.txt file, which is provided with the dbNSFP file, to obtain the correct column names. The Readme file contains also a short description of the provided values and the version of the tools used to generate them.

For more details, see here.


Enable with --vep_loftee.

For more details, see here.


Enable with --vep_spliceai. The following parameters are mandatory:

  • --spliceai_snv, to specify the path to SpliceAI raw scores snv file.
  • --spliceai_snv_tbi, to specify the path to SpliceAI raw scores snv tabix indexed file.
  • --spliceai_indel, to specify the path to SpliceAI raw scores indel file.
  • --spliceai_indel_tbi, to specify the path to SpliceAI raw scores indel tabix indexed file.

For more details, see here.


Enable with --vep_spliceregion.

For more details, see here and here.”

Plots for SnpEff are missing

When plots are missing, it is possible that the fasta and the custom SnpEff database are not matching The SnpEff completes without throwing an error causing nextflow to complete successfully. An indication for the error are these lines in the .command files:

ERRORS: Some errors were detected
Error type      Number of errors


Sentieon is a commercial solution to process genomics data with high computing efficiency, fast turnaround time, exceptional high accuracy, and 100% consistency.

In particular, Sentieon contains what may be view as speedup version of some standard GATK tools, like bwamem and haplotyper. Sarek contains support for some of the functions in Sentieon. In order to use those functions, the user will need to supply Sarek with a license for Sentieon.

Setup of Sentieon license

Sentieon supply license in the form of a string-value (a url) or a file. It should be base64-encoded and stored in a nextflow secret named SENTIEON_LICENSE_BASE64. If a license string (url) is supplied, then the nextflow secret should be set like this:

nextflow secrets set SENTIEON_LICENSE_BASE64 $(echo -n <sentieon_license_string> | base64 -w 0)

If a license file is supplied, then the nextflow secret should be set like this:

nextflow secrets set SENTIEON_LICENSE_BASE64 \$(cat <sentieon_license_file.lic> | base64 -w 0)

Available Sentieon functions

Sarek contains the following Sentieon functions bwa mem, LocusCollector + Dedup, Haplotyper, GVCFtyper and VarCal + ApplyVarCal, so the basic processing of alignment of fastq-files to VCF-files can be done using speedup Sentieon functions.

Basic usage of Sentieon functions

To use Sentieon’s aligner bwa mem, set the aligner option sentieon-bwamem. (This can, for example, be done by adding --aligner sentieon-bwamem to the nextflow run command.)

To use Sentieon’s function Dedup, specify sentieon_dedup as one of the tools. (This can, for example, be done by adding --tools sentieon_dedup to the nextflow run command.)

To use Sentieon’s function Haplotyper, specify sentieon_haplotyper as one of the tools. This can, for example, be done by adding --tools sentieon_haplotyper to the nextflow run command. In order to skip the GATK-based variant-filter, one may add --skip_tools haplotyper_filter to the nextflow run command. Sarek also provides the option sentieon_haplotyper_emit_mode which can be used to set the emit-mode of Sentieon’s haplotyper. Sentieon’s haplotyper can output both a vcf-file and a gvcf-file in the same run; this is achieved by setting sentieon_haplotyper_emit_mode to <vcf_emit_mode>,gvcf, where <vcf_emit_mode> is variant, confident or all.

To use Sentieon’s function GVCFtyper along with Sention’s version of VQSR (VarCal and ApplyVarCal) for joint-germline genotyping, specify sentieon_haplotyper as one of the tools, set the option sentieon_haplotyper_emit_mode to gvcf, and add the option joint_germline. This can, for example, be done by adding --tools sentieon_haplotyper --joint_germline --sentieon_haplotyper_emit_mode gvcf to the nextflow run command.

Joint germline variant calling

Sentieon’s GVCFtyper does not support the GenomicsDB datastore format. This means that, in contrast to the GATK based joint germline variant calling subworkflow in Sarek, the Sentieon/DNAseq based joint germline variant calling subworkflow does not use the GenomicsDB datastore format.

QualCal (BQSR)

Currently, Sentieon’s version of BQSR, QualCal, is not available in Sarek. Recent Illumina sequencers tend to provide well-calibrated BQs, so BQSR may not provide much benefit. By default Sarek runs GATK’s BQSR; that can be skipped by adding the option --skip_tools baserecalibrator.

Requested resources for the tools

Resource requests are difficult to generalize and are often dependent on input data size. Currently, the number of cpus and memory requested by default were adapted from tests on 5 ICGC paired whole-genome sequencing samples with approximately 40X and 80X depth. For targeted data analysis, this is overshooting by a lot. In this case resources for each process can be limited by either setting --max_memory and -max_cpus or tailoring the request by process name as described here. If you are using sarek for a certain data type regulary, and would like to make these requests available to others on your system, an institution-specific, pipeline-specific config file can be added here.