holtgrewe/ngs_pipelines

Nextflow-based pipelines for typical NGS processing tasks

NGS Pipelines

Getting Started

First, install the dependencies:

• Nextflow (http://nextflow.io)
• FastQC (http://www.bioinformatics.babraham.ac.uk/projects/fastqc/)
• Skewer (https://github.com/relipmoc/skewer)
• BWA (https://github.com/lh3/bwa)
• Samblaster (https://github.com/lh3/bwa)
• Samtools (http://www.htslib.org/)
• STAR aligner (https://github.com/alexdobin/STAR)
• Freebayes (https://github.com/ekg/freebayes)
• htslib (http://www.htslib.org/)
• vcflib (https://github.com/ekg/vcflib)
• bedtools2 (https://github.com/arq5x/bedtools2)
• GNU parallel (should have a package in your Linux distribution)

Then, clone the repository:

# git clone https://github.com/bihealth/ngs_pipelines.git

And run the pipelines

# nextflow run align_dna.nf \
    --dataDir $HOME/Data/2015_03_11_triple_cohort/wes_tumor \
    --runID MM065_wes_tumor \
    --runPlatform Illumina \
    -resume \
    -with-trace
# nextflow run align_dna.nf \
    --dataDir $HOME/Data/2015_03_11_triple_cohort/wes_blood \
    --runID MM065_wes_blood \
    --runPlatform Illumina \
    -resume \
    -with-trace
# nextflow run align_rna.nf \
    --runID MM065_rna_tumor \
    --runPlatform Illumina \
    --dataDir $HOME/Data/2015_03_11_triple_cohort/rna_tumor \
    -resume \
    -with-trace
# nextflow run call_multisample.nf \
    --dataDir $HOME/Data/2015_03_11_triple_cohort/variant_calling \
    --inputBam $HOME/Data/2015_03_11_triple_cohort/wes_tumor/bam/MM065_wes_tumor.bam:$HOME/Data/2015_03_11_triple_cohort/wes_blood/bam/MM065_wes_blood.bam \
    --poolID MM065_wes_tumor.MM065_wes_blood \
    -resume \
    -with-trace

Directory Structure

The pipeline expects to have one project directory (2015_03_11_triple_cohort in the example above) with one sub-directory for each sequenced sample (wes_blood, wes_tumor, and rna_tumor).

For the input, each sample folder should have a subdirectory fastq/original in which the original FASTQ files reside. Currently, only paired reads are supported, the left reads ("first read in pair") should have a name matching the pattern *_1.fastq.gz or *_R1.fastq.gz. The second read should then have the name ${NAME}_2.fastq.gz or ${NAME}_R2.fastq.gz where ${NAME} is the prefix of the first read. There can be multiple read pairs in the input directory.

SAMPLE_OR_WETLAB_ID
`-- fastq
    `-- original

After running both the alignment and the variant calling, the resulting folder structure will look as follows:

2015_03_11_triple_cohort
|-- rna_tumor
|   |-- bam
|   |-- fastq
|   |   |-- original
|   |   `-- trimmed
|   `-- reports
|       |-- alignment
|       |-- fastqc-original
|       |-- fastqc-trimmed
|       |-- split_at_n
|       `-- trimming
|-- variant_calling
|   `-- vcf
|-- wes_blood
|   |-- bam
|   |-- fastq
|   |   |-- original
|   |   `-- trimmed
|   `-- reports
|       |-- fastqc-original
|       |-- fastqc-trimmed
|       `-- trimming
`-- wes_tumor
    |-- bam
    |-- fastq
    |   |-- original
    |   `-- trimmed
    `-- reports
        |-- fastqc-original
        |-- fastqc-trimmed
        `-- trimming

For each sample, the following subdirectories exist:

• fastq/trimmed FASTQ files after adapter trimming.
• bam Aligned reads in BAM format.
• reports Logs from the trimming/alignment and QC reports.

Also, there is the folder variant_calling in the project main directory that contains a vcf folder that has the overall calls and calls filtered down to the UCSC and CCDS exons.