coVar

coVar is a tool for detecting physically-linked mutations in wastewater genomic sequencing data. Given a sorted, indexed BAM file, reference genome and gene annotation, coVar identifies and counts sequencing reads with unique physically linked mutations.

coVar also ships with Freyja, and can be accessed via the freyja covariants subcommand.

Installation

Bioconda (recommended)

conda create -n covar-env
conda activate covar-env
conda install -c bioconda covar
covar --version

crates.io (recommended)

cargo install covar
covar --version

Local build from source (experimental)

git clone https://github.com/andersen-lab/covar.git
cd covar
cargo install --path .
covar --version

Usage

covar --input <INPUT_BAM> --reference <REFERENCE_FASTA> --annotation <ANNOTATION_GFF>

Required arguments

Flag	Description
-i, --input <INPUT_BAM>	Input BAM file (must be primer trimmed, sorted, and indexed).
-r, --reference <REFERENCE_FASTA>	Reference genome in FASTA format.
-a, --annotation <ANNOTATION_GFF>	Annotation GFF3 file for translating nucleotide to amino acid mutations.

Optional arguments

Flag	Default	Description
-o, --output <OUTPUT>	stdout	Output file path. If not provided, results will be printed to stdout.
-s, --start_site <START>	0	Genomic start position for variant calling.
-e, --end_site <END>	reference length	Genomic end position for variant calling. Defaults to the length of the reference genome.
-d, --min_depth <DEPTH>	1	Minimum coverage depth for a mutation cluster to be considered.
-f, --min_frequency <FREQ>	0.001	Minimum mutation frequency (cluster depth / total depth).
-q, --min_quality <QUAL>	20	Minimum base quality score for variant calling.
-t, --threads <THREADS>	1	Number of threads to use for processing.

Example Commands

Basic run

covar \
  -i sample.bam \
  -r reference.fasta \
  -a annotation.gff3

Specify genomic region and output file

covar \
  -i sample.bam \
  -r reference.fasta \
  -a annotation.gff3 \
  -s 1000 \
  -e 5000 \
  -o output.tsv

Multi-threaded run with custom depth, quality and frequency thresholds

covar \
  -i sample.bam \
  -r reference.fasta \
  -a annotation.gff3 \
  -d 5 \
  -q 30 \
  -f 0.01 \
  -t 4

Output

The output is a tab-delimited file (.tsv) with the following columns:

Column	Description
nt_mutations	Nucleotide mutations for this cluster
aa_mutations	Corresponding amino acid translations (where possible)
cluster_depth	Total number of read pairs with this cluster of mutations
total_depth	Total number of reads spanning this cluster
frequency	Mutation frequency (cluster depth / total depth)
coverage_start	Maximum read start site for which this cluster was detected
coverage_end	Minimum read end site for which this cluster was detected

Note: Not all nucleotide mutations will have a corresponding amino acid mutations. For example, SNPs in codons that span reads or frameshift indels will be translated as 'Unknown' and 'NA', respectively.

Additionally, in the case of multiple nucleotide mutations in a single codon, the amino acid translation will occur multiple times. For example, if a codon has two SNPs, the amino acid translation will be repeated for each SNP. See the following example:

nt_mutations:
G22813T T22882G G22895C T22896C G22898A A22948C A22948-CCT
aa_mutations:
S:K417N S:N440K S:V445P S:V445P S:G446S S:K462N S:DEL463

Here, the SNPS G22895C and T22896C are in the same codon, which result in Proline (P) at position 445. In order to keep a one-to-one relationship between nucleotide and amino acid mutations, the amino acid translation will be repeated so we see S:V445P twice.