deepvariant_germline

1
# This command assumes all the inputs are in the current working directory and all the outputs go to the same place.
2
docker run --rm --gpus all --volume $(pwd):/workdir --volume $(pwd):/outputdir \
3
    --workdir /workdir \
4
    nvcr.io/nvidia/clara/clara-parabricks:4.7.1-1 \
5
    pbrun deepvariant_germline \
6
    --ref /workdir/${REFERENCE_FILE} \
7
    --in-fq /workdir/${INPUT_FASTQ_1} /workdir/${INPUT_FASTQ_2} \
8
    --out-variants /outputdir/${OUTPUT_VCF_FILE} \
9
    --out-bam /outputdir/${OUTPUT_BAM_FILE}

1
# Run bwa-mem and pipe output to create sorted BAM
2
$ bwa mem \
3
    -t 32 \
4
    -K 10000000 \
5
    -R '@RG\tID:sample_rg1\tLB:lib1\tPL:bar\tSM:sample\tPU:sample_rg1' \
6
    <INPUT_DIR>/${REFERENCE_FILE} \
7
    <INPUT_DIR>/${INPUT_FASTQ_1} <INPUT_DIR>/${INPUT_FASTQ_2} | \
8
  gatk SortSam \
9
      --java-options -Xmx30g \
10
      --MAX_RECORDS_IN_RAM 5000000 \
11
      -I /dev/stdin \
12
      -O cpu.bam \
13
      --SORT_ORDER coordinate
14
15
# Mark Duplicates
16
$ gatk MarkDuplicates \
17
    --java-options -Xmx30g \
18
    -I cpu.bam \
19
    -O mark_dups_cpu.bam \
20
    -M metrics.txt
21
22
# Run deepvariant
23
BIN_VERSION="1.9.0"
24
sudo docker run \
25
  -v "${PWD}":"/input" \
26
  -v "${PWD}/output":"/output" \
27
  -v "${PWD}/Ref":"/reference" \
28
  google/deepvariant:"${BIN_VERSION}" \
29
  /opt/deepvariant/bin/run_deepvariant \
30
  --model_type WGS \
31
  --ref /reference/Homo_sapiens_assembly38.fasta \
32
  --reads /output/${OUTPUT_BAM_FILE} \
33
  --output_vcf /output/"${OUTPUT_VCF_FILE}" \
34
  --num_shards $(nproc) \
35
  --disable_small_model=true \
36
  --make_examples_extra_args "ws_use_window_selector_model=true"