For those who wish to map their metatranscriptome(s) to GTDB, I can recommend downloading the

ncbi_genome_infos

and Sourmash indexes prior to running the pipeline (via

<https://ftp.ncbi.nlm.nih.gov/genomes/ASSEMBLY_REPORTS/assembly_summary_refseq.txt>

and

<https://ftp.ncbi.nlm.nih.gov/genomes/ASSEMBLY_REPORTS/assembly_summary_genbank.txt>

). Sourmash GTDB release 226:

<https://farm.cse.ucdavis.edu/~ctbrown/sourmash-db/gtdb-rs226/>

My

params.yaml

contained the following relevant parameters:

input: 'samplesheet.csv'
outdir: 'output'
sequence_filter: './data/SILVA_138.1_allrRNAref_NR99_tax.fasta'
indexes: './sourmash/gtdb-rs226-reps.k31.sig.zip'
sourmash: true
ksize: 31
ncbi_genome_infos: './ncbi_genome_infos.csv'

In my case,

BBMAP_ALIGN

and

BBMAP_INDEX

required additional CPUs, so `nextflow.config`:

process {
    withName: BBMAP_ALIGN {
        memory = 128.GB
        cpus = 36
        time = 10.h
    }
    withName: BBMAP_INDEX {
        memory = 72.GB
        cpus = 24
        time = 10.h
    }
}

So the directory from where I launched nf-core/magmap contained six files:

assembly_summary_genbank.txt

,

assembly_summary_refseq.txt

,

ncbi_genome_infos.csv

,

nextflow.config

,

params.yaml

, and

samplesheet.csv

Command:

nextflow run nfcore/magmap -r dev -resume -params-file params.yaml -profile pdc_kth

Workflow:

nf-core/magmap: v1.0.0-g36313b8

Nextflow: 25.04.3

Thank you Daniel for all the help!

And thank you @George Westmeijer for this! I'll try to remember to get the relevant portions of this into the documentation before 1.0 release!