LGvar: Large-scale Genomic VARiation

Quick Start

• Why LGvar?
• Installation
• Usage & Examples
  • Reverse-complement the genome
  • Generate orthologous chromosome pairs
  • Whole genome SV identification
  • Output files
  • Plot whole genome alignment synteny
• Benchmarking
• Getting Help
• Citing LGvar

Why LGvar

LGvar is a comprehensive toolset designed for large-scale structural variant (SV) detection based on genome assemblies. It excels in cross-species variant identification, particularly demonstrating superior performance in inversion detection compared to existing tools.

• Easily use: Only FASTA and chromosome pairs (.tsv) are needed for input (users can alternatively provide pre-aligned PAF files).
• Comprehensive and fast: LGvar detects not only insertions, deletions, inversions, translocations, duplications, INDELs and SNVs, but also detects structurally divergent regions (SDRs) and nested-inversions (INV-INV).
• More efficient in detection of inversions: LGvar performs realignment within SDRs to detect inversions missed in initial alignment.

Installation

Option 1: Conda (Recommended)

# You can install LGvar from anaconda.org
conda install -c conda-forge -c bioconda lgvar

# Or you can directly create a new environment
# Clone the repository  
git clone https://github.com/YafeiMaoLab/LGvar.git 
cd LGvar
# Create and activate environment  
conda env create -f env.yml  
conda activate LGvar

Option 2: Singularity

# If conda is not available, use our pre-built images:
singularity pull library://feifeizhou/tool/lgvar:v1.3.0 
# Execution example:  
singularity exec lgvar_v1.3.0.sif /LGVAR/LGVAR [subcommands]

Usage & Examples

• LGvar provides several subcommands to streamline the genomic analysis pipeline:

  Subcommands List (click to expand)

  Command	Function
  run	Run SV identification. Detects SDRs, DELs, INSs, INVs, DUPs, INDELs, SNVs etc.
  plot	Syntenic Plot. Visualizes the alignment between genomes. Blue regions represent syntenic blocks, while orange regions represent inversions.
  pair	Homology Mapping. Generates homologous chromosome pair files.
  reverse	RC-Check. Identifies and reverse-complements misoriented chromosomes.

1. Reverse-complement the genome

• Before calling SVs, using LGvar reverse to check if query chromosomes need reverse-complementing to ensure inversion calling accuracy, this step identifies a chromosome as needing reverse complementation by checking if the total length of alignments mapped to the reverse strand exceeds half the chromosome’s total length in the PAF file.

  # Recommended: align first then check orientation  
  LGVAR reverse -p align.paf -r reverse.txt -g original.genome.fa -n new.genome.fa
  # Or use the script
  bash /src/scripts/rc_chrom.sh align.paf reverse.txt original.genome.fa new.genome.fa
  #Then use new.genome.fa as query genome to do SV calling
  

  reverse.txt contains the chromosomes that need to be reverse-complemented.

2. Generate orthologous chromosome pairs

• LGvar identifies variants between each pair of orthologous chromosomes. If you do not have a tab-delimited TSV file containing ref query columns, you can generate the orthologous chromosome TSV file using the LGvar pair command.

  LGVAR pair -p align.paf -l 1000000 -o pair
  

  where the -l parameter specifies the minimum alignment length (in base pairs) required to define a reliable orthologous chromosome pair; only alignments meeting or exceeding this 1Mbp threshold will be used to determine the corresponding relationships between chromosomes. It is recommended to increase this value when analyzing two species with a large evolutionary divergence (e.g. human vs. lemur, -l can be set to 5Mbp).

3. Whole genome SV identification

• Once you have prepared the reference genome, query genome and orthologous chromosome pair files, you can use LGVAR run to perform genome-wide variant detection. You may optionally provide files containing centromere and telomere coordinates of the reference genome to mask highly repetitive regions, thus improving running efficiency.

  LGVAR run \
      -r ref.fa \
      -q1 query.hap1.fa \
      -q2 query.hap2.fa \
      -cp1 query.hap1.pair.tsv \
      -cp2 query.hap2.pair.tsv \
      -m sensitive \
      -s query
  

  Centromere and telomere coordinate files for T2T-CHM13 are available under the examples/data directory, namely chm13_cen.tsv and chm13_telo.tsv.

• Here is an running example. The test data is located in examples/. We'll use Human chromosome 21(T2T-CHM13) as reference and Chimpanzee (PTR) as query.

  # 1. Setup working directory  
  mkdir LGvar_work
  # 2. Decompress test genomes  
  cd examples/genome && gunzip *.gz  
  cd LGvar_work
  # 3. Execution (Example with existing PAF alignments)  
  LGVAR run \  
      -r /examples/genome/chr21.chm13.fa \ 
      -q1 /examples/genome/chr21.ptr.hap1.fa \  
      -q2 /examples/genome/chr21.ptr.hap2.fa \  
      -p1 /examples/align/align_hap1.paf \  
      -p2 /examples/align/align_hap2.paf \
      -cp1 /examples/data/PTR_hap1_pairs.tsv \
      -cp2 /examples/data/PTR_hap2_pairs.tsv \
      -m sensitive \
      -s PTR
  

4. Output Files

• Results are saved in the ${work_dir}/results folder:

  results

  File	Function
  sortLGvar_all.vcf	Combined results of SNVs, INDELs and SVs for both haplotypes.
  LGvarhap1(2).vcf	Haplotype-specific SV details.
  LGvarhap1(2).bed	Haplotype-specific SV details (with SDRs, DUPs, TRANs).
  LGvar.bed	Merged variant regions (Hap1 + Hap2).

  Label	Description
  SNVs	Single Nucleotide Variation
  INDELs	Small insertions and deletions (Length < 50bp)
  DELs/INSs	Deletions and Insertions (filtered at 50bp threshold)
  INVs/INV-INV	Simple and Nested Inversions
  DUP	Duplications
  TRANS	Translocations
  SDRs	Structurally Divergent Regions

5. Plot whole genome alignment synteny

• You can visualize whole-genome alignments using PAF and PAIR files.

  LGVAR plot -p align.paf -f pair.tsv -o align.pdf
  

  This is an example using T2T-CHM13 as reference and chimpanzee hap1 as query genome.

Benchmarking

• We benchmarked LGvar in many genomes include simulated, population and cross-species data (see forthcoming paper). See Analysis.md:

Getting help

• If you encounter any issues or have questions about specific parameters, please Open an Issue or contact zhoufeifei@sjtu.edu.cn.

Citing LGvar

• If you use LGvar in your research, please cite our repository (and forthcoming paper): (https://github.com/YafeiMaoLab/LGvar)