FAQ and support

Processing large numbers of genomes

If you plan to process a large number of genomes, you may wish to break these into smaller batches. With CheckM 1.3.0, running 1,000 genomes at a time with 64 threads takes ~20 minutes and 12 GB of memory. Exceeding the available memory of your machine will cause CheckM to use swap space (as per any program), which will substantially increase the time to process genomes; use --gplace_max_memory to limit the memory used to place genomes in the reference genome tree (see System requirements). CheckM’s output files are gzip-compressed (see Output files), and --rm_tmp_files can be used to remove the intermediate files of a workflow once it has finished.

It has also been reported that processing large numbers of genomes at once can cause a Python recursion error. As such, splitting large runs into batches of 1,000 genomes is recommended.

How do I improve the quality estimates for CPR genomes?

Use the CPR marker set provided with CheckM, as described in Using the CPR marker set.

Where is the CheckM reference data?

See Reference data for how to download the reference data and tell CheckM where it is located.

Reporting a bug

When submitting a bug, please provide as much information as possible and, if feasible, a simplified version of the data set which is causing the issue. The more information you provide, the quicker we will be able to resolve the issue. Please report bugs through the GitHub issues system. The log file checkm.log written to the output directory is useful to include with a bug report.

Known issues are tracked in the GitHub issues.

Feature requests

CheckM v1 is no longer actively developed: outside of minor updates to fix compatibility issues with external dependencies, no maintenance is intended. New features are not planned. CheckM2 is recommended for new analyses.

Contact information

Inquiries can be sent to Donovan Parks (donovan.parks [at] gmail.com).

Contributors

The following people have directly submitted enhancements to the CheckM code base or other substantial contributions:

Massive thanks to baudrly, Vini Salazar, and Asaf Peer for the initial Python 2 to 3 porting of CheckM. Other contributions are acknowledged in the Change log.