rust-read-extraction 0.1.0

Fast BAM to FASTQ conversion with region filtering for SDrecall

rust-read-extraction

A fast Rust-based BAM to FASTQ converter with region filtering, designed for the SDrecall workflow.

Features

• Fast BAM to FASTQ conversion: Efficiently extracts paired-end reads from BAM files
• Region-based filtering: Extracts only reads overlapping specified BED regions
• Multi-aligned read filtering: Optional filtering for multi-aligned reads based on alignment tags (SA, XA, AS, XS)
• Paired-end aware: Ensures both mates are extracted when one overlaps the target region

Installation

pip install rust_read_extraction

Usage

from rust_read_extraction import bam_to_fastq_biobambam

# Extract reads from specific regions
r1_path, r2_path = bam_to_fastq_biobambam(
    input_bam="aligned.bam",
    region_bed="regions.bed",
    output_freads="output_R1.fastq",
    output_rreads="output_R2.fastq",
    multi_aligned=False,  # Set True to filter multi-aligned reads
    threads=4,
    tmp_dir="/tmp"
)

Parameters

• input_bam: Path to indexed BAM file
• region_bed: BED file specifying regions of interest
• output_freads: Output path for R1 FASTQ
• output_rreads: Output path for R2 FASTQ
• multi_aligned: Whether to apply multi-aligned read filtering (default: False)
• threads: Number of threads for BAM reading (default: 1)
• tmp_dir: Temporary directory (default: "/tmp")

Multi-aligned Read Filtering

When multi_aligned=True, the module filters reads based on:

• MAPQ < 60 (mandatory)
• No SA tag (no supplementary alignments)
• Has XA tag (has secondary alignments)
• |AS - XS| <= 10 (alignment score difference)

License

MIT License

Author

Xingtian Yang ([e-mail])