Skip to content

Latest commit

 

History

History
99 lines (70 loc) · 2.83 KB

svanalyzer.md

File metadata and controls

99 lines (70 loc) · 2.83 KB

SVanalyzer Quick Guide: Benchmarking and Merging Structural Variants

This guide provides essential commands for using SVanalyzer's benchmark and merge functions to analyze structural variants (SVs).

Benchmarking Structural Variants

Compare test SVs against a truth set to estimate sensitivity and specificity.

Usage

svanalyzer benchmark --ref <reference.fasta> --test <test.vcf> --truth <truth.vcf>

Key Options

  • --ref: Reference FASTA file (required).
  • --test: VCF file with test variants (required).
  • --truth: VCF file with true variants (required).
  • --maxdist: Maximum positional distance between variants to consider a match (default: 100,000).
  • --normshift: Maximum normalized shift allowed (default: 0.2).
  • --normsizediff: Maximum normalized size difference (default: 0.2).
  • --normdist: Maximum normalized edit distance (default: 0.2).
  • --minsize: Minimum variant size to include (default: 0).
  • --prefix: Prefix for output files (default: benchmark).

Example

svanalyzer benchmark \
  --ref reference.fasta \
  --test sample_test.vcf \
  --truth sample_truth.vcf \
  --prefix benchmark_results

Merging Structural Variants

Group SVs by similarity to create a unified set of variants.

Usage

Merge from a single VCF file:

svanalyzer merge --ref <reference.fasta> --variants <variants.vcf> --prefix <output_prefix>

Merge from multiple VCF files:

svanalyzer merge --ref <reference.fasta> --fof <vcf_list.txt> --prefix <output_prefix>

Key Options

  • --ref: Reference FASTA file.
  • --variants: VCF file containing variants to merge.
  • --fof: File listing paths to VCF files (one per line).
  • --prefix: Prefix for output files (default: merged).
  • --maxdist: Maximum distance between variants for merging (default: 2000).
  • --reldist: Maximum normalized edit distance (default: 0.2).
  • --relsizediff: Maximum normalized size difference (default: 0.2).
  • --relshift: Maximum normalized shift (default: 0.2).
  • --seqspecific: Only include variants with explicit REF and ALT sequences.

Example

Merging from multiple VCF files:

  1. Create a file listing your VCFs (e.g., vcf_list.txt):

    caller1.vcf
caller2.vcf
caller3.vcf

  2. Run the merge command:

    svanalyzer merge \
  --ref reference.fasta \
  --fof vcf_list.txt \
  --prefix merged_results

Notes

  • Reference Consistency: Ensure the reference FASTA file matches the one used for generating your VCFs.
  • Parameter Tuning: Adjust --maxdist, --reldist, etc., based on your data characteristics.
  • Sequence-Specific Variants: Use --seqspecific to focus on variants with defined REF and ALT sequences.
  • Output Files: The --prefix option sets the prefix for all output files generated.