recombass 0.1.6

Detect recombination hot/cold spots from SNP matrices using wavelet denoising

recombass

A command-line tool for detecting recombination hotspots and coldspots from bacterial SNP matrices using wavelet-based denoising and PMR (Pairwise Mismatch Rate) analysis.

Designed for microbial population genomics, recombass integrates strain filtering, MAF pruning, distance computation, and signal smoothing to identify genomic regions with elevated or suppressed recombination signals.

Requirements

• Python 3.9+
• snp-dists available on PATH

Installation

From PyPI

Install the external runtime dependency first, then install recombass from PyPI:

mamba install -c conda-forge -c bioconda snp-dists
pip install recombass

Basic Usage

Input Format

The input SNP matrix must be a tab-delimited table with:

• SNP positions in the first column
• strain names in the header row
• nucleotide calls in each cell

Example:

	strain1	strain2	strain3	strain4
100	A	T	A	G
200	C	C	T	C
300	G	A	G	A
400	T	T	C	T
500	G	G	A	G

Command Line Interface

The main command structure:

recombass [OPTIONS] -m  -i <input_path> -o <output_prefix>

Available Options

• -m, --maf FLOAT: Max major allele frequency (e.g., 0.95) (highly suggested)
• -n, --nonredundant FLOAT: Non-redundant strain clustering threshold (e.g., 0.01)
• --cold-criterion FLOAT: Coldspot detection threshold (default: -0.4)
• --hot-criterion FLOAT: Hotspot detection threshold (default: 0.6)
• -t, --threads INT: Number of threads to use (default: 20)
• -i, --input PATH: Input file path (required)
• -o, --output PATH: Output prefix (required)

Example Usage

# Basic usage
recombass -i <snps.tsv> -o <result_prefix>

# With MAF filtering and non-redundant filtering
recombass -i <snps.tsv> -o <result_prefix> -m 0.95 -n 0.01

# With custom thresholds and threads
recombass -i <snps.tsv> -o <result_prefix> -m 0.9 -n 0.02 --cold-criterion -0.5 --hot-criterion 0.7 -t 10

Output Files

The pipeline can generate the following files:

• <result_prefix>.nr{rate}: Non-redundant filtered SNP matrix when -n is used
• <result_prefix>.maf{rate}: MAF-filtered SNP matrix when -m is used
• <result_prefix>.fa: FASTA converted from the current SNP matrix
• <result_prefix>.fa.dist: Pairwise SNP distance matrix from snp-dists
• <result_prefix>.fa.dist.tr: Triangular distance table used downstream
• <result_prefix>.fa.dist.png: Distance distribution plot
• <result_prefix>.recx.txt: Recombination index value
• <result_prefix>.fa.pmr.30.2.wt.l.pdf: Hotspot plot from the denoised close-strain PMR signal
• <result_prefix>.fa.pmr.all.2.wt.l.pdf: Coldspot plot from the denoised all-strain PMR signal
• <result_prefix>.integrated.tsv: Integrated per-site output with raw PMR, denoised PMR, and hot/cold labels

When filtering is enabled, the effective output prefix changes as intermediate files are generated. For example:

recombass -i sample.tsv -o sample_out -m 0.95 -n 0.01

will produce final files such as:

• sample_out.nr01
• sample_out.nr01.maf95
• sample_out.nr01.maf95.integrated.tsv
• sample_out.nr01.maf95.fa.pmr.30.2.wt.l.pdf
• sample_out.nr01.maf95.fa.pmr.all.2.wt.l.pdf

License

This project is licensed under the MIT License.