snakemake 命令

常用示例

说明

Snakemake is a workflow management system for creating reproducible and scalable data analyses. Workflows are defined using a Python-based domain-specific language in files called "Snakefiles" that describe rules for transforming input files into output files. Snakemake automatically determines the execution order by building a directed acyclic graph (DAG) of jobs based on file dependencies. It only re-executes rules when input files are newer than outputs, avoiding redundant computations. Workflows can seamlessly scale from single workstations to clusters and cloud environments without modification. The tool is particularly popular in bioinformatics and data science for its ability to handle complex pipelines with many interdependent steps. It supports conda environments, containers (Docker/Singularity), and various cluster execution backends.

参数

-n, --dry-run

Show the execution plan without actually running any jobs.

-p, --printshellcmds

Print shell commands that will be executed.

--cores, -c _N_

Use at most N CPU cores in parallel. If N is omitted, uses all available cores.

-s, --snakefile _FILE_

Specify the Snakefile. Default: Snakefile in current directory.

-d, --directory _DIR_

Execute workflow in the specified directory.

--forceall, -F

Force re-execution of all rules.

--forcerun, -R _rules_

Force re-execution of specific rules.

--until, -U _rules_

Run only until the specified rules.

--dag

Output the directed acyclic graph of jobs in DOT format.

--rulegraph

Output the dependency graph of rules in DOT format.

--config _key=value_

Set or override config values.

--configfile _FILE_

Specify a config file in YAML or JSON format.

--profile _PROFILE_

Use a workflow profile for execution settings.

--cluster _CMD_

Execute jobs on a cluster using the given submit command.

--use-conda

Use conda environments defined in rules.

--use-singularity

Use Singularity containers defined in rules.

-q, --quiet

Suppress output except warnings and errors.

--help

Display help information.