3. Long-Read Structural Variant (SV) Discovery & Phasing
Uncovers large deletions, insertions, inversions, duplications, translocations, and complex chromosomal rearrangements (>50 bp) using PacBio HiFi or Oxford Nanopore reads with Sniffles2, CuteSV, and SURVIVOR.
| Metric / Requirement | Specification | Target Configuration |
|---|---|---|
| Core Engine / Framework | Minimap2 + Sniffles2 + CuteSV + SURVIVOR | Nextflow / Snakemake Compatible |
| Benchmark Runtime | ~4 – 6 hours (30× PacBio HiFi / ONT WGS) | 32 CPU cores / 64 GB RAM |
| Standard Container | quay.io/biocontainers/sniffles:2.3.3--pyhdfd78af_0 | Docker • Singularity • Conda |
| Input → Deliverable | HiFi / ONT FASTQ → Phased Multi-Caller Structural Variant VCF | |
Step 1: Base-Accuracy Long-Read Alignment with Minimap2
Aligns continuous long reads against GRCh38 using chained seed matching optimized for high-accuracy PacBio HiFi reads or long ONT reads.
# Map PacBio HiFi reads (or change -ax to map-ont for Nanopore)
minimap2 -ax map-hifi -t 32 --MD -Y \
-R "@RG\tID:HIFI_01\tSM:SAMPLE_01\tPL:PACBIO" \
/references/GRCh38.fa \
hifi_reads.fastq.gz | \
samtools sort -@ 8 -m 4G -o hifi_aligned.bam -
samtools index -@ 8 hifi_aligned.bam-Y enables soft clipping for supplementary alignments; --MD calculates mismatch and deletion strings critical for structural indel callers.Step 2: Structural Variant Discovery & Haplotype Phasing with Sniffles2
Detects SVs (>50 bp) directly from split-read and CIGAR indel evidence while resolving heterozygous calls into maternal and paternal haplotypes.
sniffles \
--input hifi_aligned.bam \
--vcf hifi_sniffles_sv.vcf.gz \
--reference /references/GRCh38.fa \
--threads 32 \
--minsvlen 50 \
--phase \
--output-rnames--minsvlen 50 focuses calling on genuine structural variants; --phase performs inline phasing of SV alleles across heterologous long reads.Step 3: Orthogonal Structural Variant Calling with CuteSV
Deploys multi-tier clustering and refinement of signatures to identify breakpoint positions for deletions, insertions, duplications, and inversions.
mkdir -p cutesv_work/
cuteSV hifi_aligned.bam /references/GRCh38.fa hifi_cutesv.vcf cutesv_work/ \
--threads 32 \
--min_size 50 \
--min_support 3 \
--genotype \
--max_cluster_bias_INS 100 \
--diff_ratio_filtering_INS 0.3 \
--max_cluster_bias_DEL 100 \
--diff_ratio_filtering_DEL 0.3
bgzip -c hifi_cutesv.vcf > hifi_cutesv.vcf.gz && tabix -p vcf hifi_cutesv.vcf.gz--min_support 3 ensures high sensitivity for mosaic or low-depth structural variants; --genotype computes diploid genotypes (0/1, 1/1).Step 4: Consensus SV Merging Across Calling Algorithms with SURVIVOR
Combines callsets across Sniffles2 and CuteSV within a 1,000 bp breakpoint distance window to yield high-confidence consensus structural variants.
ls hifi_sniffles_sv.vcf.gz hifi_cutesv.vcf.gz > sv_vcf_list.txt
# Parameters: list, max_distance=1000bp, min_callers=2, type_agree=1, strand_agree=1, disabled=0, min_size=50
SURVIVOR merge sv_vcf_list.txt 1000 2 1 1 0 50 merged_high_confidence_sv.vcf
bgzip -c merged_high_confidence_sv.vcf > merged_high_confidence_sv.vcf.gz
tabix -p vcf merged_high_confidence_sv.vcf.gzmin_callers=2 eliminates caller-specific single-tool false positive structural artifacts.Step 5: Structural Variant Functional Impact & Gene Disruption with AnnotSV
Annotates structural variants with affected gene models, TAD boundaries, OMIM phenotypes, pathogenic ClinVar records, and DGV population frequencies.
AnnotSV -SVinputFile merged_high_confidence_sv.vcf.gz \
-outputDir ./annotsv_results/ \
-genomeBuild GRCh38 \
-annotationMode full-annotationMode full classifies structural variants by both full transcript span and promoter/exon junction breakpoint impacts.🚀 Automated Nextflow (nf-core) Production Execution
Execute this complete workflow across any SLURM, AWS Batch, or Google Cloud HPC cluster using the verified community container pipeline:
nextflow run nf-core/structuralvariant -r 1.0.0 \
--input samplesheet.csv \
--tools sniffles,cutesv \
--genome GRCh38 \
--outdir ./results_sv/ \
-profile docker