covfast

A fast, C-based utility for calculating template-based coverage from CRAM/BAM files. covfast calculates depth of coverage statistics for specified genomic regions. It is designed to calculate template coverage, meaning that for a given read pair, any overlapping bases between the two reads are counted only once. This provides a measure of unique fragment coverage, which can be more accurate for certain applications than the standard alignment-based coverage. The program's logic and filtering are designed to closely match the output of samtools mpileup.

Key Features: - Template-Based Coverage: Avoids double-counting overlapping regions of read pairs. - High Performance: Written in C with htslib for efficient CRAM/BAM processing. - Comparable to samtools: Mimics the filtering logic of samtools depth -s for comparable results. - Flexible Filtering: Allows for custom setting of minimum mapping quality (MAPQ) and base quality.

Build Instructions:

Requirements: gcc make wget Standard htslib dependencies: zlib, bzip2, lzma, curl, openssl. Compiling: The included Makefile will automatically download and compile the required version of htslib before building the covfast executable. Build the executable: make

This will create the covfast binary in the current directory.

Install to system (Optional):

To install the covfast executable to /usr/local/bin for system-wide access, run: sudo make install Usage

Usage: ./covfast [OPTIONS]

Positional Arguments: A standard BED file defining regions of interest (0-based). An indexed CRAM or BAM file with read alignments.

Options: -r, --reference The reference genome FASTA file, required for CRAM. -o, --outfile Output file for statistics. [Default: stdout] -q, --minmapqual Minimum mapping quality. [Default: 1] -Q, --minbasequal Minimum base quality. [Default: 13] -c, --coverage_values Comma-separated coverage thresholds for summary stats. -h, --help Show this help message and exit.

Example:

covfast --reference /path/to/hg38.fa \ --outfile coverageresults.tsv \ /path/to/myregions.bed \ /path/to/sample.cram

Docker:

For a fully containerized and reproducible environment, you can use the provided Dockerfile. See the Docker README for detailed instructions.