MGnify genomes generation pipeline (GGP) produces prokaryotic and eukaryotic MAGs from raw reads and corresponding assemblies.
This pipeline does not support co-binning.
The pipeline performs the following tasks:
- Supports short reads.
- Changes read headers to their corresponding assembly accessions (in the ERZ namespace).
- Quality trims the reads, removes adapters fastp.
Afterward, the pipeline:
- Runs a decontamination step using BWA to remove any host reads. By default, it uses the hg38 human genome reference.
- Bins the contigs using Concoct, MetaBAT2 and MaxBin2.
For prokaryotes:
- Refines the bins using Binette.
- Conducts bin quality control with CAT, GUNC, and CheckM.
- Performs dereplication with dRep.
- Calculates coverage using MetaBAT2 calculated depths.
- Detects rRNA and tRNA using cmsearch.
- Assigns taxonomy with GTDBtk.
For eukaryotes:
- Estimates quality and merges bins using EukCC.
- Dereplicates MAGs using dRep.
- Calculates coverage using MetaBAT2 calculated depths.
- Assesses quality with BUSCO and EukCC.
- Assigns taxonomy with BAT.
Final steps:
- Tools versions are available in software_versions.yml
- Pipeline generates a tsv table for public MAG uploader
- TODO: finish MultiQC
If this the first time running nextflow please refer to this page
You need to download the mentioned databases and specify as inputs to parameters.
| Tool/Database | Version | Purpose |
|---|---|---|
| BUSCO | 5.4.7 (DB v2024-01-08) | Assign genome quality |
| CAT | 5.2.3 (DB v2021-01-07) | Taxonomic classification |
| CheckM2 | 1.0.1 | Determining genome quality |
| EukCC | 2.1.3 (DB v1.2) | Completeness and contamination of eukaryotic genomes |
| GUNC | 4 (DB v2.0.4) | Quality control |
| GTDB-Tk + ar53_metadata_r*.tsv, bac120_metadata_r*.tsv from here | 2.3.0 (DB release214) | Assigning taxonomy; generating alignments |
| Rfam | 14.9 | Database for identification of SSU/LSU rRNA and other ncRNAs |
| Human reference genome hg38 | hg38 | The reference genome of your choice for decontamination including bwa-mem2 index, for example, human. Format: genome.fna; bwa-mem2/ folder containing: - *.fna.0123- *.fna.amb- *.fna.ann- *.fna.bwt.2bit.64- *.fna.pac |
If you will use ENA data follow instructions. Otherwise, download your data and keep format as recommended in inputs description below.
Each row corresponds to a specific dataset with information such as an identifier for the row, the file path to the assembly (assembly), and paths to the raw reads files (fastq_1 and fastq_2). Additionally, the assembly_accession column contains associated assembly accessions.
| id | assembly | fastq_1 | fastq_2 | assembly_accession |
|---|---|---|---|---|
| SRR1631112 | /path/to/ERZ1031893.fasta | /path/to/SRR1631112_1.fastq.gz | /path/to/SRR1631112_2.fastq.gz | ERZ1031893 |
There is an example here.
ID: run accession
Assembly_software: tool that was used to assemble run into assembly.
| id | assembly_software |
|---|---|
| SRR1631112 | Assembler_vVersion |
Manually choose the most appropriate metagenome from https://www.ebi.ac.uk/ena/browser/view/408169?show=tax-tree.
For example, marine metagenome
Comma-separated environment parameters in format:
"environment_biome,environment_feature,environment_material"
For example, marine sediments,subtropical gyre,sinking marine particle
nextflow run ebi-metagenomics/genomes-generation \
-profile `specify profile(s)` \
--input `samplesheet.csv` \
--assembly_software_file `software.tsv` \
--metagenome "chosen metagenome" \
--biomes "chosen biome,chosen feature,chosen material" \
--outdir `full path to output directory`--xlarge (default=false): use high-memory config for big studies. Study maybe considered as big if it has more than 300 runs. In addition, if study has less number of runs but they are very deeply sequenced it also makes sense to try that option.--skip_preprocessing_input (default=false): skip input data pre-processing step that renames ERZ-fasta files to ERR-run accessions. Useful if you process data not from ENA--skip_decontamination (default=false): skip decontamination on reference genome--skip_prok (default=false): do not generate prokaryotic MAGs--skip_euk (default=false): do not generate eukaryotic MAGs--merge_pairs (default=false): merge paired-end reads on QC step with fastp
Use final_table_for_uploader.tsv to upload your MAGs to ENA with uploader.
Example of final_table_for_uploader.tsv.
! Do not modify existing output structure because that TSV file contains full paths to your genomes.
For a more detailed description of the different output files, see the outputs file.
If you use this pipeline please make sure to cite all used software.