VACmap

VACmap-—a long-read aligner specifically designed for complex structural variation discovery.

Installation

git clone https://github.com/micahvista/VACmap.git
cd VACmap
conda env create --name vacmap_env --file VACmap_environment.yml
conda activate vacmap_env
python setup.py install

Usage

Map long genomic reads:
---------------------- 

vacmap -ref /ref.fasta -read /read.fasta -mode S -t 8 | samtools sort -@4 > alignments.sorted.bam
samtools index -@4 alignments.sorted.bam

*Memory usage: <20GB
*Speed: Processing HG002 50X ONT data (170GB) took 5.8 hours using 40 threads.

Map assembly:
---------------------- 

vacmap -ref /ref.fasta -read /read.fasta -mode asm -t 8  -workdir /home/usr/workdir/ --H --fakecigar | samtools sort -@4 > alignments.sorted.bam
samtools index -@4 alignments.sorted.bam

or (lower memory usage)

vacmap -ref /ref.fasta -read /read.fasta -mode asm -t 8  -workdir /home/usr/workdir/ --H --fakecigar > alignments.sam
samtools sort -@4 alignments.sam > alignments.sorted.bam
samtools index -@4 alignments.sorted.bam

*Memory usage: 30GB-70GB Using more threads requires more memory.
*Speed: Processing HG002 v1.0 assembly(48 contigs) took 2.3 hours using 40 threads.

Parameter

Mandatory parameter:

    -ref The path of reference sequence. 
    -read The path of long reads. 
    -t The number of threads to use. (Note: In asm mode, using fewer threads results in lower memory usage.)
    -mode S|L|H|R|asm
    -workdir Only for asm mode, store temporary data. If the folder not existexist, VACmap will create it automatically.

Input:
    Supported read input types: FASTA, FASTA.GZ, FASTQ, FASTQ.GZ, BAM.
        [Warning: Duplicate reads (based on read names) will be processed only once.]
    Multiple read input options:
        -read read_1.fa read_2.fa ...
        -read read_1.fa -read read_2.fa ...

Mapping:
    -mode H For aligning high error rate long read (Pacbio CLR, ONT). 
    -mode L For aligning low error rate long read (Pacbio HiFi). 
    -mode S Increase the sensitivity for small variants. (<100bp). (Pacbio CLR, ONT, HiFi). 
    -mode R Use a fixed value for the variation penalty, more sensitive to translocation events, 
        such as gene conversion. (Pacbio CLR, ONT, HiFi). 
    -mode asm For full genome alignment.   
    
    -k k-mer size (no larger than 28, deflaut: 15) # set -k 19 -w 10 for HiFi data 
        to reduce run time (2X faster) but there is very small decrease in accuracy.
    -w minimizer window size. (deflaut: 10)


Output: 

    --eqx Output =/X CIGAR operators for sequence match/mismatch.
    
    --MD Output the MD tag.
    
    --cs[=short|long] Output the cs tag. (deflaut: short cs).
    
    --copycomments Copy FASTA/Q comments to output. 

    --H Use hard-clipping instead of soft-clipping for clipped sequences.
        [Warning: SV callers may fail to detect split-read events.]

    --Q Ignore base quality in the input file.

    --fakecigar Use approximate CIGAR in the SA tag.

    --rg-id <string> Adds RG:Z:<string> to all alignments
    --rg-sm <string> RG header: Sample 
    --rg-lb <string> RG header: Library 
    --rg-pl <string> RG header: Platform
    --rg-ds <string> RG header: Description
    --rg-dt <string> RG header: Date (format: YYYY-MM-DD)
    --rg-pu <string> RG header: Platform unit 
    --rg-pi <string> RG header: Median insert size
    --rg-pg <string> RG header: Programs 
    --rg-cn <string> RG header: sequencing center
    --rg-fo <string> RG header: Flow order 
    --rg-ks <string> RG header: Key sequence 
    --rg-pm <string> Platform model. Free-form text providing further details of the platform/technology used.
    --rg-bc <string> Barcode sequence identifying the sample or library.

Limitations

Single Reference Sequence Size Limit: 
    2,147,483,647 bases.
Single Query Sequence Size Limit:
    2,147,483,647 bases for Modes L, H, S, R.
    Unlimited for Mode asm.

Demo

This is a demonstration of the non-linear alignment algorithm used in VACmap: http://154.12.33.179:8050/. I hope this tool proves helpful for your research!

Changelog

Full genome alignment. Oct 3, 2024

Reduce runtime (45% faster for aligning HiFi data) and include options to decrease output size. Seq 3, 2024

Reduce runtime (45% faster for aligning HiFi data) and include options to decrease output size. Seq 3, 2024

Copy FASTA/Q comments to output. Aug 22, 2024

Improved speed, 40% faster. Aug 9, 2024

Fix bug. Aug 5, 2024

Enabled overlapping anchors in non-linear chaining. Jul 4, 2024

TODO

Reimplement VACmap in C:
    Seed: Completed
    Non-linear Chaining: Completed
    Extend: Completed
    Testing

Contact

If you experience any problems or have suggestions please create an issue or a pull request.

Citation

https://www.biorxiv.org/content/10.1101/2023.08.03.551566v3

License

The project is licensed under the GNU General Public License.