- align HiFi reads to reference with pbmm2
- call small variants with DeepVariant v1.4
- phase small variants with WhatsHap
- haplotag aligned BAMs with WhatsHap and merge
.
├── cluster_logs # slurm stderr/stdout logs
├── reference
│ ├── reference.chr_lengths.txt # cut -f1,2 reference.fasta > reference.chr_lengths.txt
│ ├── reference.fasta
│ └── reference.fasta.fai
├── samples
│ └── <sample_id> # sample_id regex: r'[A-Za-z0-9_-]+'
│ ├── whatshap/ # phased small variants; merged haplotagged alignments
│ ├── logs/ # per-rule stdout/stderr logs
│ ├── aligned/ # intermediate
│ ├── deepvariant/ # intermediate
│ ├── deepvariant_intermediate/ # intermediate
│ └── whatshap_intermediate/ # intermediate
├── smrtcells
│ ├── done # move folders from smrtcells/ready to smrtcells/done to prevent re-processing
│ └── ready
│ └── <sample_id> # uBAMs or FASTQs per sample
│ # filename regex: r'm\d{5}[Ue]?_\d{6}_\d{6}).(ccs|hifi_reads).bam' or r'm\d{5}[Ue]?_\d{6}_\d{6}).fastq.gz'
└── workflow # clone of this repo
# create the base conda environment
$ conda create \
--channel bioconda \
--channel conda-forge \
--prefix ./conda_env \
python=3 snakemake=6 mamba lockfile
# activate the base conda environment
$ conda activate ./conda_env
# clone the github repo
$ git clone https://github.com/williamrowell/hifi-deepvariant-snakemake.git workflow
# create a few directories
$ mkdir -p reference smrtcells/{ready,done} cluster_logs
# drop your reference.fasta and reference.fasta.fai into reference and adjust the path in workflow/config.yaml
# copy or link some data into a smrtcells/ready/<sample_id> directory
$ ln -s /path/to/some/hifi/data/m00000_000000_000000.hifi_reads.bam smrtcells/ready/<sample_id>/m00000_000000_000000.hifi_reads.bam
$ ln -s /path/to/some/hifi/data/m11111_111111_111111.hifi_reads.bam smrtcells/ready/<sample_id>/m11111_111111_111111.hifi_reads.bam
# run the workflow for sample <sample_id>
$ sbatch workflow/run_snakemake.sh <sample_id>