checkm2_predict

Tags: quality-control completeness contamination machine-learning bacteria archaea sample-scope

Assess genome quality using machine learning.

Uses CheckM2 to predict the completeness and contamination of genome assemblies. Unlike the original CheckM, it uses a gradient boost machine learning model to predict quality without relying on lineage-specific marker sets, making it more accurate for novel or reduced genomes.

Database Required

Requires the CheckM2 database (Diamond database file) to be available.

Inputs

record (
    meta: Record,
    fna: Path
)

Field	Type	Description
meta	Record	Groovy Record containing sample information
fna	Path	Assembled contigs in FASTA format

db: Path

Name	Type	Description
db	Path	The CheckM2 database file (*.dmnd)

Outputs

record (
    meta: Record,
    tsv: Path,
    results: Set<Path>,
    logs: Set<Path?>,
    nf_logs: Set<Path>,
    versions: Set<Path>
)

Field	Type	Description
meta	Record	Sample information record
tsv	Path	Tab-delimited report of quality metrics (Completeness, Contamination)
results	Set<Path>	All output files to be published
logs	Set<Path?>	Optional program specific log files
nf_logs	Set<Path>	Nextflow-specific log files (e.g. .command.{begin
versions	Set<Path>	A YAML formatted file with program versions

Parameters

CheckM2 Parameters

Parameter	Type	Default	Description
--checkm2_lowmem	boolean		Low memory mode. Reduces DIAMOND blocksize to significantly reduce RAM usage at the expense of longer runtime
--checkm2_general	boolean		Force the use of the general quality prediction model (gradient boost)
--checkm2_specific	boolean		Force the use of the specific quality prediction model (neural network)
--checkm2_allmodels	boolean		Output quality prediction for both models for each genome.
--checkm2_genes	boolean		Treat input files as protein files. [Default: False]
--checkm2_opts	string		Additional options to pass to CheckM2

Used By

Subworkflows

• checkm2 - Assess metagenome bin completeness using CheckM2.

Workflows

• checkm2 - Machine learning-based assessment of microbial genome assembly quality.

Citations

If you use this in your analysis, please cite the following.

• Bactopia
  Petit III RA, Read TD Bactopia - a flexible pipeline for complete analysis of bacterial genomes. mSystems 5 (2020)

• CheckM2
  Chklovksi A Rapid assessment of genome bin quality using machine learning (GitHub)

Source

View source on GitHub

Version

CHECKM2_PREDICT:
    - checkm2: 1.1.0