Intro

The Nextflow scripts in this repository were mostly used for a specific H3Africa project. The project contained datasets at different stages e.g Fastq, BAM and gVCF so it made sense just to process the set at stages where necessary.

To run

Each folder contains part of the pipeline or separate scripts to prepare the data for a specific part of the pipeline or get it ready for archival purposes.

Main pipeline

• align - Need to provide a sample sheet containing paths to forward and reverse reads. BWA, Picard MarkDuplicates and GATK BaseRecalibrator are run on the samples.
• generate-gvcf - Need to provide a sample sheet with paths to the BAMs. GATKs HaplotypeCaller was run in gVCF mode on the samples.
• combine-gvcfs - Need to provide a sample sheet with paths to the gVCFs. GATKs CombineGVCFs were run on all samples.
• gene-calling - Need to provide the path to the combined gVCF. GenotypeGVCFs are run jointly on all samples on gene regions only.
• genome-calling - Need to provide the path to the combined gVCF. GenotypeGVCFs are run jointly on all samples on full genome. VQSR is also applied to chromosome 1 to 22, X, Y and MT. GenotypeGVCFs are run on chromosome level where SNP and INDEL VQSR are run on genome level (according to GATKs best practices).
• genome-calling-vqsr-per-chromosome Need to provide the path to the combined gVCF. GenotypeGVCFs are run jointly on all samples on full genome. VQSR is also applied to chromosome 1 to 22. Calling is only done for X, Y and MT in a separate Nextflow script. GenotypeGVCFs are run on chromosome level where SNP and INDEL VQSR are run on chromosome level.

Separate pipelines

• bam-to-cram - Need to provide a sample sheet with paths to the BAMs. BAMs are converted to CRAM (v3), indexed, stats are calculated and md5sums are generated.
• cram-to-fastq - Need to provide a sample sheet with paths to the CRAMs. CRAMs are converted to Fastq (forward and reverse pair).`
• index-bams - Need to provide a sample sheet with paths to the BAMs. BAMs are indexed.`
• select-from-combined-gvcf - Need to provide the path to the combined GVCF and list of sample ids to be removed. Sample ids are removed from the combined gVCF and a new combined gVCF are generated.`

Note

• The main pipeline can be followed for a single sample or joint calling sample. For a single sample the combine-gvcf can be skipped and the gene-calling and genome-calling configs can point to single sample .g.vcf.

Example

To get test data that van be used from the alignment to calling and VQSR phase download the GiaB dataset here

wget -O NA12878_R1.fastq.gz ftp://ftp-trace.ncbi.nih.gov/giab/ftp/data/NA12878/Garvan_NA12878_HG001_HiSeq_Exome/NIST7035_TAAGGCGA_L001_R1_001.fastq.gz
wget -O NA12878_R2.fastq.gz ftp://ftp-trace.ncbi.nih.gov/giab/ftp/data/NA12878/Garvan_NA12878_HG001_HiSeq_Exome/NIST7035_TAAGGCGA_L001_R2_001.fastq.gz

To do

1. Pull all software as Docker containers (tabix, samtools, GATK).
2. Currently setup for b37. Setup a swith in nextflow to allow for processing against build b38. This would required switching between chromosome names and reference databases
3. All output alignments in the align step should be in CRAM