Skip to content

Latest commit

 

History

History
146 lines (116 loc) · 5.93 KB

installation.md

File metadata and controls

146 lines (116 loc) · 5.93 KB

Installation

Before you can run TOFU-MAaPO, you need to install Singularity and Nextflow. We show how to install the dependencies in the the Quick start section or you install them manually by following the links.

You will need to prepare databases for respective modules (e.g., Metaphlan4, HUMAnN3) and a config file for your compute system.

Configuration

Before running TOFU-MAaPO for the first time, you must set up a configuration file tailored to your computing system. There are two approaches for configuration:

Option 1: Modify the repository's config file (not recommended)

  • Edit the custom.config file located in the repository's conf directory.
  • Run the pipeline using the custom profile: 
    -profile custom

Caution: This approach is less flexible and may cause issues when updating the repository, as your changes might be overwritten.

Option 2: Create a separate configuration file (recommended)

  • Create and maintain a separate configuration file (e.g., tofu.config).
  • Include your custom configuration in the pipeline execution by adding: 
    -profile custom -c /path/to/tofu.config

In this example tofu.config should contain following (from you customized) lines: 

params {
  //Software DB locations, UNCOMMENT AND CHANGE THEM:

  //metaphlan_db = "${baseDir}/databases/Metaphlan/4.0"
  //kraken2_db = "${baseDir}/databases/Kraken2/k2_viral_20210517"
  //humann_db = "${baseDir}/databases/Humann3/3.6" 
  //gtdbtk_reference = "${baseDir}/databases/GTDB-TK/release207_v2"

  //For host read removal list your host genomes as bowtie2 index in this named list with full path to the basename of the index:
	'genomes' {
		'human' {
      bowtie_index = false
		}
	}

  // MAXIMUM PER PROCESS CONFIGS, CHANGE THEM TO YOUR HARDWARE SPECS
  max_memory = 100.GB
  max_cpus = 32
  max_time = 48.h
  
  //Scratch. Does your system support scratch? If not, leave it false
  scratch = false
}

//Enable Singularity as container software
singularity {
	enabled = true
  // Singularity configs, CHANGE THEM TO YOUR USED FILESYSTEM, if not properly set the container won't see your files
	runOptions = "-B /home -B /tmp"
	// where should the containers be downloaded to
	cacheDir = "${launchDir}/singularity_cache"
}

process {
  //Default executor for each process, other options can be e.g. SLURM.
  //See https://www.nextflow.io/docs/latest/executor.html for more options and details.
  executor='local'
}

//Default for total execution, remove this whole part if you are using a different option above than executor='local':
executor {
  cpus = 32
  memory = 200.GB
}

Executors

If you intend to run the pipeline on an HPC or cloud service, refer to the Nextflow Executors documentation for guidance on adapting your custom configuration file to your environment.

QC: Host Decontamination

To perform host decontamination, you need the appropriate host genome Bowtie2 indexes. Download these indexes, for example, from Bowtie2 AWS indexes, and specify the path to the basename of the index files in your custom configuration file before running the pipeline. Add the following snippet to your configuration file:

'genomes' {
    'human' {
        bowtie_index = "/path/to/your/references/iGenomes/references/Homo_sapiens/NCBI/GRCh38Decoy/Sequence/Bowtie2Index/genome"
    }
}

MetaPhlAn & HUMAnN

The pipeline requires databases to run HUMAnN and MetaPhlAn.

Use the pipeline's built-in download functionality for HUMAnN and MetaPhlAn databases with the parameters:

--updatemetaphlan for MetaPhlAn4.
--updatehumann for HUMAnN3.

Supply the paths to store the databases to using:
--metaphlan_db for MetaPhlAn4.
--humann_db for HUMAnN3.

Run the pipeline with these parameters during your initial run.
Ensure your local computer has internet access, as the pipeline will automatically download and extract the databases.

Hint: Once the first run is successful, update your custom configuration file to permanently set the paths for --metaphlan_db and --humann_db.

Note: Running HUMAnN3 also requires the MetaPhlAn4 database (version vJan21).

Update your configuration file with the following snippet:

params {
    metaphlan_db = "/path/to/your/databases/Metaphlan/4.0"
    humann_db = "/path/to/your/databases/Humann3/3.6"
}

Kraken2

Download and extract a Kraken2 database, such as those available from Kraken2 AWS indexes. You can specify the database path either in your custom configuration file or dynamically during each run with the parameter --kraken2_db.

Note: If you plan to use Bracken, ensure the Kraken2 database is Bracken-ready.

Add the following snippet to your custom configuration file:

params {
    kraken2_db = "/path/to/your/databases/Kraken2/krakendb"
}

Genome Assembly

For genome assembly, the pipeline requires GTDB-Tk reference data (currently Release R207_v2). Use the pipeline's built-in download functionality to download the GTDBTk database with the parameter --updategtdbtk and specify the path where to store the database with --gtdbtk_reference.

Add the following snippet to your custom configuration file:

params {
	gtdbtk_reference = "/path/to/your/databases/GTDB-TK/release207_v2"
}

Kiel Medcluster

On Kiel Medcluster, please load the following modules with:

module load singularity nextflow

When starting TOFU-MAaPO, make sure the medcluster profile is selected, by using: -profile medcluster No further action is required!
Databases for all tools (with the exception of Salmon) and the host genomes human, mouse and chimp are already set up.