nf-core/mag

Use this to specify the location of your input FastQ files and their associated metadata. You can also use the CSV file to assign different groups or to include long reads for hybrid assembly with metaSPAdes. The CSV file must have at least two columns (sample, short_reads1) and with a maximum CSV sheet having the headers: sample,run,group,short_reads_1,short_reads_2,long_reads. See usage docs.

By default, the pipeline expects paired-end data. If you have single-end data, you need to specify --single_end on the command line when you launch the pipeline. A normal glob pattern, enclosed in quotation marks, can then be used for --input. For example:

--single_end --input '*.fastq'  

It is not possible to run a mixture of single-end and paired-end files in one run.

If you have pre-computed assemblies from another source, it is possible to jump straight to the binning stage of the pipeline by supplying these assemblies in a CSV file. This CSV file should have at minimum three columns and the following header: id,group,assembler,fasta (group is only required when --coassemble_group). Short reads must still be supplied in to --input` in CSV format. See usage docs for further details.

Set this parameter to your e-mail address to get a summary e-mail with details of the run sent to you when the workflow exits. If set in your user config file (~/.nextflow/config) then you don't need to specify this on the command line for every run.

MEGAHIT only generates reproducible results when run single-threaded.

When using this parameter do not change the number of CPUs for the megahit process with a custom config file. This would result in an error.

Default: The number of CPUs is specified in the base.config file, and increased with each retry.

SPAdes is designed to be deterministic for a given number of threads. To generate reproducible results fix the number of CPUs using this parameter.

When using this parameter do not change the number of CPUs for the spades process with a custom config file. This would result in an error.

Default: -1 (the number of CPUs is specified in the base.config or in a custom config file, and increased with each retry).

SPAdes is designed to be deterministic for a given number of threads. To generate reproducible results fix the number of CPUs using this parameter.

When using this parameter do not change the number of CPUs for the spadeshybrid process with a custom config file. This would result in an error.

Default: -1 (the number of CPUs is specified in the base.config or in a custom config file, and increased with each retry).

MetaBAT2 is run by default with a fixed seed within this pipeline, thus producing reproducible results. You can set it also to any other positive integer to ensure reproducibility. Set the parameter to 0 to use a random seed.

Default base-quality trimming is set to trim by 'windows', as in FastP. Specifying this flag will use trim via contiguous stretch of low quality bases (Ns) instead.

Replaces --trimwindows 4 with --trimqualities in AdapterRemoval

This parameter is mutually exclusive with --host_fasta. Host read removal is done with Bowtie2.
Both the iGenomes FASTA file as well as corresponding, already pre-built Bowtie 2 index files will be used.

This parameter is mutually exclusive with --host_genome. The reference can be masked. Host read removal is done with Bowtie2.

This parameter must be used in combination with --host_fasta, and should be a directory containing files from the output of bowtie2-build, i.e. files ending in .bt2

The default value focuses on length instead of quality to improve assembly size.
In order to assign equal weights to read lengths and read qualities set this parameter to 1.
This might be useful, for example, to benefit indirectly from the removal of short host reads (causing lower qualities for reads not overlapping filtered short reads).

Local directory containing *.cf files, or a URL or local path to a downloaded compressed tar archive of a Centrifuge database. E.g. ftp://ftp.ccb.jhu.edu/pub/infphilo/centrifuge/data/p_compressed+h+v.tar.gz.

Path to a local directory, archive file, or a URL to compressed tar archive that contains at least the three files hash.k2d, opts.k2d and taxo.k2d. E.g. ftp://ftp.ccb.jhu.edu/pub/data/kraken2_dbs/minikraken_8GB_202003.tgz.

Path to taxonomy.tab file for Krona, instead of downloading the default file. Point at the .tab file.

E.g. https://tbb.bio.uu.nl/bastiaan/CAT_prepare/CAT_prepare_20210107.tar.gz. This parameter is mutually exclusive with --cat_db_generate. The file needs to contain a folder named *taxonomy* and *database* that hold the respective files.

Download the taxonomy files from NCBI taxonomy, the nr database and generate CAT database. This parameter is mutually exclusive with --cat_db. Useful to build a CAT database with the same DIAMOND version as used for running CAT classification, avoiding compatibility problems.

Useful to allow reproducibility, as old versions of prebuild CAT databases do not always remain accessible and underlying NCBI taxonomy and nr databases change.

Completeness assessed with BUSCO analysis (100% - %Missing). Must be greater than 0 (min. 0.01) to avoid GTDB-tk errors. If too low, GTDB-tk classification results can be impaired due to not enough marker genes!

Contamination approximated based on BUSCO analysis (%Complete and duplicated). If too high, GTDB-tk classification results can be impaired due to contamination!

A low number of CPUs helps to reduce the memory required/reported by GTDB-Tk. See also the GTDB-Tk documentation.

Will be faster than writing to disk (default setting), however at the expense of much larger memory (RAM) requirements for GDTBTK/CLASSIFY.

Must be a directory containing the uncompressed contents from https://zenodo.org/doi/10.5281/zenodo.6994741 (nf-core/mag tested with v1.1)

An example is adjusting k-mers ("-k 21,33,55,77") or adding advanced options. But not --meta, -t, -m, -o or --out-prefix, because these are already in use. Must be used like this: --spades_options "-k 21,33,55,77")

An example is adjusting presets (e.g. "--presets meta-large"), k-mers (e.g. "-k 21,33,55,77") or adding other advanced options. For example, increase the minimum k-mer in the event of an error message such as "Too many vertices in the unitig graph, you may increase the kmer size to remove tons of erroneous kmers." in the MEGAHIT log file. But not --threads, --memory, -o or input read files, because these are already in use. Must be used like this: --megahit_options "--presets meta-large"

Sometimes Prokka will complain that your contig names are too long and fail.

This particularly happens with metaSPAdes assemblies.

You can turn on this flag which will tell Prokka to truncate the contig names for you.
However this also requires you to specify a sequencing centre name (specified with --prokka_compliance_centre).

:::warning
Truncating contig names may make it harder to associated contig annotations with their original contigs!
:::

Specify the sequencing centre name for making NCBI Genbank/ENA compatible annotation files (required when specifying --prokka_with_compliance).

mmseqs2 lists a large number of databases, not all of which are appropriate for use with MetaEuk. MetaEuk requires protein inputs, so you should select one of the Aminoacid or Profile options.

One option would be the databases from the MetaEuk publication (https://wwwuser.gwdg.de/~compbiol/metaeuk/), however it should be noted that these are focused on marine eukaryotes.

Available: all, group or own. Note that own cannot be specified in combination with --coassemble_group.

Note that specifying all without additionally specifying --coassemble_group results in n^2 mapping processes for each assembly method, where n is the number of samples.

For forwarding into downstream analysis, i.e. QUAST and BUSCO, and reporting.

Contigs that do not fulfill the thresholds of --min_length_unbinned_contigs and --max_unbinned_contigs are pooled for downstream analysis and reporting, except contigs that also do not fullfill --min_contig_size are not considered further.

Contigs that do not fulfill the thresholds of --min_length_unbinned_contigs and --max_unbinned_contigs are pooled for downstream analysis and reporting, except contigs that also do not fullfill --min_contig_size are not considered further.

Bowtie2 alignment mode options, for example: --very-fast , --very-sensitive-local -N 1 , ... Must be used like this: --bowtie2_mode "--very-sensitive"

Specify to save the BAM and BAI files generated when mapping input reads back to the assembled contigs (performed in preparation for binning and contig depth estimations).

Enable this if it is likely that your metagenome samples contain a mixture of eukaryotic and prokaryotic genomes. This will ensure that prokaryote-only steps only receive putatively prokaryotic genomes, and vice-versa. Additionally, may improve the performance of DAS Tool by ensuring it only receives prokaryotic genomes.

E.g. https://busco-data.ezlab.org/v5/data/lineages/bacteria_odb10.2024-01-08.tar.gz or '/path/to/buscodb' (files still need to be unpacked manually). Available databases are listed here: https://busco-data.ezlab.org/v5/data/lineages/.

Useful to allow reproducibility, as BUSCO datasets are frequently updated and old versions do not always remain accessible.

By default, BUSCO creates a large number of intermediate files every run. This may cause problems on some clusters which have file number limits in plate, particularly with large numbers of bins. Enabling this option cleans these files, reducing the total file count of the work directory.

The pipeline can also download this for you if not specified, and you can save the resulting directory into your output directory by specifying --save_checkm_data. You should move this directory to somewhere else on your machine (and supply back to the pipeline in future runs again with --checkm_db.

If specified, the directories and files decompressed from the tar.gz file downloaded from the CheckM FTP server will be stored in your output directory alongside your CheckM results.

The pipeline can also download this for you if not specified, and you can save the resulting directory into your output directory by specifying --save_checkm2_data. You should move this directory to somewhere else on your machine (and supply back to the pipeline in future runs again with --checkm2_db).

If specified, the directories and files decompressed from the tar.gz file downloaded from the Zenodo repository will be stored in your output directory alongside your CheckM2 results.

Score threshold for single-copy gene selection algorithm to keep selecting bins, with a value ranging from 0-1.

For description of scoring algorithm, see: Sieber, Christian M. K., et al. 2018. Nature Microbiology 3 (7): 836–43. https://doi.org/10.1038/s41564-018-0171-1.

Modifies DAS Tool parameter --score_threshold

raw_bins_only: only bins (and unbinned contigs) from the binners.
refined_bins_only: only bins (and unbinned contigs) from the bin refinement step .

~~both: bins and unbinned contigs from both the binning and bin refinement steps.~~ both option is disabled in v2.4 due a bug that will be fixed in a later release.

If specified, the corresponding DIAMOND file downloaded from the GUNC server will be stored in your output directory alongside your GUNC results.