Skip to content

Latest commit

 

History

History
72 lines (43 loc) · 3.38 KB

config.md

File metadata and controls

72 lines (43 loc) · 3.38 KB

2.2 Configuration files

Pipeline settings are stored in configuration files for convenience, see /confs folder for examples. A .config file can be passed to the nextflow executor via the -c argument.

For starter, copy one of the example config files in /confs and modify it dependent on your analysis requirements. Then run the pipeline by including your config file.

nextflow run RNASeq-NF/main.nf -c </path/to/your/settings.config> --fastq_path <path/to/fastq/files> --out_dir </path/to/output_dir> -profile slurm -resume

Executors and logs

Nextflow's base configuration (executors, containerization etc.) should always be included in the analysis config. Set the cacheDir to a location with enough space to store your singularity images.

cacheDir = '/path/to/my/cache/dir

Pipeline parameters

Analysis parameters, such as specified in the params section, can be stored either in the the config file (recommended) or appended on the command line, for example --singleEnd.

Genome configuration

In it's most basic form, the pipeline requires the following resource parameters.

  • --genome_fasta - path to genome sequence (.fasta).
  • --genome_gtf - path to genome annotation (.gtf)

Transcript quantification with Salmon requires transcript sequences in .fasta format to be supplied (In case no Salmon index is provided)

  • --transcripts_fasta - path to transcript sequences (.fasta).

GATK requires a sequence dictionary (.dict) and fasta index (.fai) to be present in the same directionary as the reference sequence. Please prepare them before running the pipeline if applicable as described in step 2 and 3 of the GATK reference preparation guide

Known variants in vcf format (optional) should be included in the genome config as shown below.

 genome_known_sites = ['/hpc/cog_bioinf/common_dbs/GATK_bundle/1000G_phase1.indels.b37.vcf',
  '/hpc/cog_bioinf/common_dbs/GATK_bundle/dbsnp_137.b37.vcf',
  '/hpc/cog_bioinf/common_dbs/GATK_bundle/Mills_and_1000G_gold_standard.indels.b37.vcf']

All remaining files will be created at run-time and stored in the output folder (reference_resources). Since buidling the STAR index requires a significant amount of memory, ensure that sufficient memory is available on the host system and allocated in process.config.

Note

Generating resource files at run-time should only be done upon running the pipeline for the first time with a specific genome build. Include the resource files in your genome config for subsequent analysis to avoid re-building and putting unnecessary strains on the host-system. Example genome configuration files can be found in ./resources/test_run.

Process configuration

Runtime specific resources (memory, cpu's) should be sufficent for most jobs but can be always be altered if required.

  withLabel : SortMeRNA_4_2_0 {
      time = '24h'
      penv = 'threaded'
      cpus = 4
      memory = '15G'
      publishDir.path = "${params.out_dir}/QC/"
      publishDir.mode = 'copy'
      publishDir.saveAs = {filename ->
                     if (filename.indexOf("_rRNA_report.txt") > 0) "SorteMeRNA/$filename"
                     else if (filename.indexOf("_filtered_rRNA.fastq.gz") > 0) "SorteMeRNA/rRNA-reads/$filename"
                     else null }

  }