Bio Variant Calling Deepvariant

Singularity

singularity pull docker://google/deepvariant:1.6.1

Basic Usage

Goal: Call germline variants from aligned reads using DeepVariant's deep learning model.

Approach: Run the all-in-one run_deepvariant wrapper specifying model type, reference, reads, and output paths.

"Call variants with DeepVariant" → Convert aligned read pileups into image tensors, classify with a CNN, and output genotyped VCF.

One-Step Run (run_deepvariant)

docker run -v "${PWD}:/input" -v "${PWD}/output:/output" \
    google/deepvariant:1.6.1 \
    /opt/deepvariant/bin/run_deepvariant \
    --model_type=WGS \
    --ref=/input/reference.fa \
    --reads=/input/sample.bam \
    --output_vcf=/output/sample.vcf.gz \
    --output_gvcf=/output/sample.g.vcf.gz \
    --num_shards=16

Model Types

Step-by-Step Workflow

Goal: Run DeepVariant in three explicit stages for more control over intermediate outputs.

Approach: Generate pileup image tensors (make_examples), classify with the CNN (call_variants), then merge and genotype (postprocess_variants).

For more control, run each step separately:

Step 1: Make Examples

docker run -v "${PWD}:/data" google/deepvariant:1.6.1 \
    /opt/deepvariant/bin/make_examples \
    --mode calling \
    --ref /data/reference.fa \
    --reads /data/sample.bam \
    --examples /data/examples.tfrecord.gz \
    --gvcf /data/gvcf.tfrecord.gz

Step 2: Call Variants

docker run -v "${PWD}:/data" google/deepvariant:1.6.1 \
    /opt/deepvariant/bin/call_variants \
    --outfile /data/call_variants.tfrecord.gz \
    --examples /data/examples.tfrecord.gz \
    --checkpoint /opt/models/wgs/model.ckpt

Step 3: Postprocess Variants

docker run -v "${PWD}:/data" google/deepvariant:1.6.1 \
    /opt/deepvariant/bin/postprocess_variants \
    --ref /data/reference.fa \
    --infile /data/call_variants.tfrecord.gz \
    --outfile /data/output.vcf.gz \
    --gvcf_outfile /data/output.g.vcf.gz \
    --nonvariant_site_tfrecord_path /data/gvcf.tfrecord.gz

GPU Acceleration

Goal: Speed up DeepVariant inference using GPU hardware.

Approach: Use the GPU-enabled container image with Docker --gpus flag.

docker run --gpus all -v "${PWD}:/data" \
    google/deepvariant:1.6.1-gpu \
    /opt/deepvariant/bin/run_deepvariant \
    --model_type=WGS \
    --ref=/data/reference.fa \
    --reads=/data/sample.bam \
    --output_vcf=/data/output.vcf.gz \
    --num_shards=16

PacBio HiFi Calling

Goal: Call variants from PacBio HiFi long reads.

Approach: Use the PACBIO model type which is trained on HiFi read characteristics.

docker run -v "${PWD}:/data" google/deepvariant:1.6.1 \
    /opt/deepvariant/bin/run_deepvariant \
    --model_type=PACBIO \
    --ref=/data/reference.fa \
    --reads=/data/hifi_aligned.bam \
    --output_vcf=/data/hifi_variants.vcf.gz \
    --num_shards=16

ONT Calling

Goal: Call variants from Oxford Nanopore long reads.

Approach: Use the ONT_R104 model type trained on Nanopore R10.4 chemistry.

docker run -v "${PWD}:/data" google/deepvariant:1.6.1 \
    /opt/deepvariant/bin/run_deepvariant \
    --model_type=ONT_R104 \
    --ref=/data/reference.fa \
    --reads=/data/ont_aligned.bam \
    --output_vcf=/data/ont_variants.vcf.gz \
    --num_shards=16

Exome/Targeted Sequencing

Goal: Call variants from exome or targeted panel data.

Approach: Use WES model type with a BED file restricting calling to target regions.

docker run -v "${PWD}:/data" google/deepvariant:1.6.1 \
    /opt/deepvariant/bin/run_deepvariant \
    --model_type=WES \
    --ref=/data/reference.fa \
    --reads=/data/exome.bam \
    --regions=/data/targets.bed \
    --output_vcf=/data/exome_variants.vcf.gz \
    --num_shards=8

Joint Calling with GLnexus

Goal: Perform joint genotyping across a cohort from DeepVariant gVCFs.

Approach: Generate per-sample gVCFs, then merge and jointly genotype with GLnexus using a DeepVariant-specific config.

For multi-sample cohorts, use gVCFs with GLnexus:

# Generate gVCFs for each sample
for bam in *.bam; do
    sample=$(basename $bam .bam)
    docker run -v "${PWD}:/data" google/deepvariant:1.6.1 \
        /opt/deepvariant/bin/run_deepvariant \
        --model_type=WGS \
        --ref=/data/reference.fa \
        --reads=/data/$bam \
        --output_vcf=/data/${sample}.vcf.gz \
        --output_gvcf=/data/${sample}.g.vcf.gz \
        --num_shards=16
done

# Joint genotyping with GLnexus
docker run -v "${PWD}:/data" quay.io/mlin/glnexus:v1.4.1 \
    /usr/local/bin/glnexus_cli \
    --config DeepVariantWGS \
    /data/*.g.vcf.gz \
    | bcftools view - -Oz -o cohort.vcf.gz

GLnexus Configurations

Output Quality Metrics

Goal: Assess the quality of DeepVariant calls.

Approach: Generate summary statistics with bcftools stats and check Ti/Tv ratio as a quality indicator.

# Variant statistics
bcftools stats output.vcf.gz > stats.txt

# Filter by quality
bcftools view -i 'QUAL>20 && FMT/GQ>20' output.vcf.gz -Oz -o filtered.vcf.gz

# Ti/Tv ratio (expect ~2.0-2.1 for WGS)
bcftools stats output.vcf.gz | grep TSTV

Benchmarking Against Truth Set

Goal: Evaluate DeepVariant accuracy against a GIAB truth set.

Approach: Run hap.py to compute precision, recall, and F1 for SNPs and indels.

# Using hap.py for GIAB benchmarking
docker run -v "${PWD}:/data" jmcdani20/hap.py:latest \
    /opt/hap.py/bin/hap.py \
    /data/HG002_GRCh38_truth.vcf.gz \
    /data/deepvariant_output.vcf.gz \
    -r /data/reference.fa \
    -o /data/benchmark \
    --threads 16

Complete Workflow Script

Goal: Run DeepVariant end-to-end with indexing and statistics in a single script.

Approach: Wrap run_deepvariant, bcftools index, and bcftools stats in a parameterized shell script.

#!/bin/bash
set -euo pipefail

BAM=$1
REFERENCE=$2
OUTPUT_PREFIX=$3
MODEL_TYPE=${4:-WGS}
THREADS=${5:-16}

echo "=== DeepVariant: ${MODEL_TYPE} mode ==="

docker run -v "${PWD}:/data" google/deepvariant:1.6.1 \
    /opt/deepvariant/bin/run_deepvariant \
    --model_type=${MODEL_TYPE} \
    --ref=/data/${REFERENCE} \
    --reads=/data/${BAM} \
    --output_vcf=/data/${OUTPUT_PREFIX}.vcf.gz \
    --output_gvcf=/data/${OUTPUT_PREFIX}.g.vcf.gz \
    --intermediate_results_dir=/data/${OUTPUT_PREFIX}_tmp \
    --num_shards=${THREADS}

echo "=== Indexing ==="
bcftools index -t ${OUTPUT_PREFIX}.vcf.gz
bcftools index -t ${OUTPUT_PREFIX}.g.vcf.gz

echo "=== Statistics ==="
bcftools stats ${OUTPUT_PREFIX}.vcf.gz > ${OUTPUT_PREFIX}_stats.txt

echo "=== Complete ==="
echo "VCF: ${OUTPUT_PREFIX}.vcf.gz"
echo "gVCF: ${OUTPUT_PREFIX}.g.vcf.gz"

Comparison with Other Callers

Resource Requirements

Related Skills

• variant-calling/gatk-variant-calling - GATK alternative
• variant-calling/variant-calling - bcftools calling
• long-read-sequencing/clair3-variants - Long-read alternative
• variant-calling/filtering-best-practices - Post-calling filtering