Introduction

prototaxites/mgannotate is a bioinformatics pipeline that predicts genes in metagenome assemblies, annotates them, and produces a high-level summary of GO terms within them.

Pipeline summary

Currently, the pipeline performs the following:

• (optional) (co-)assembles metagenome shotgun short reads with MEGAHIT
• (optional) filters assembled contigs to remove or keep specific clades using MMSeqs
• Predicts of protein-coding genes using MetaEuk
• (optional) clustering of genes using MMSeqs
• Annotates predicted genes using eggnog-mapper
• Counts reads mapping to each gene using either HTSeq-count or CoverM
• Produces count summaries for each GO in a provided list

Pipeline diagram

Usage

The pipeline can be run with the following command:

nextflow run prototaxites/mgannotate \
    -profile <docker/singularity/podman/shifter/charliecloud/conda> \
    --reads <samplesheet_reads.csv> \
    --outdir <OUTDIR>

Input

Sequencing data

Input of sequencing data is via csv files, containing paths and associated metadata:

• --reads is a csv file with the following headers: sampleid,assemblyid,forward_reads,reverse_reads
• --assemblies is a csv file with the following headers: assemblyid,assembler,path

Where:

• forward_reads, reverse_reads, and path are the paths to the forward reads fastq, reverse reads fastq, and assembly fasta, respectively. Reverse reads are optional, but if using co-assembly, all reads for a given assemblyid must be paired or single-end only.
• assemblyid is the unique identifier of the assembly. If --assemblies is provided, coverage for each reads entry mapping to an assemblyid is estimated individually. If no --assemblies is provided, assemblyid is used to group reads for co-assembly.
• sampleid is a unique identifier for each shotgun sequencing sample.

Options for input are as follows:

• --reads only: Pipeline will perform denovo assembly of shotgun reads
• --assemblies only: Pipeline will annotate assemblies but not estimate gene or GO abundances
• --reads and --assemblies: Pipeline will annotate assemblies and estimate gene and GO abundances

Databases

The pipeline requires three databases for full functionality. If they are not provided, the steps requiring the databases will not be run.

• Taxonomic database (MMSeqs format):
  • This should be an MMSeqs database with taxonomic annotations. Can be provided either as a string with --mmseqs_tax_db with the names of one of the databases available at the MMSeqs2 documentation, in which case it is downloaded, or as a path to a pre-downloaded local database with --mmseqs_tax_db_local.
• Functional database (MMSeqs format):
  • This should be an MMSeqs database to use for gene predictions with MetaEuk. Can be provided either as a string with --mmseqs_func_db with the names of one of the databases available at the MMSeqs2 documentation, in which case it is downloaded, or as a path to a pre-downloaded local database with --mmseqs_func_db_local. If the eggnog-mapper MMSeqs2 database has been pre-downloaded, the path to this could also be provided.
• eggnog-mapper database:
  • Path to a directory containing the eggnog-mapper database. If not provided, the eggnog-mapper database will be automatically downloaded - this can be saved by supplying --save_eggnog_db. Can be downloaded following the instructions provided in the eggnog-mapper documentation. When downloading the database, make sure to enable downloading the MMSeqs2 database with the -M flag as this is what is presently use by the pipeline.

Pipeline flags

Some important pipeline flags to enable specific modes:

• --filter_contigs: Enable taxonomic filtering of assembly contigs using the taxids specified in --filter_taxon_list.
• --cluster_genes: Before annotation, take the gene predictions from all assemblies and cluster them together into one gene catalogue. Reads mapping to the genes are then counted using CoverM.
• --assemblies_are_genes: It is possible to skip gene prediction and provide a fasta file of gene sequences instead of metagenome assemblies. Must be used in conjunction with --cluster_genes as there is no GFF file to map with using HTSeq-Count.

Pipeline output

The output folder (--outdir) contains the following directories:

• assemblies/{assemblyid}: denovo assembly fasta files
• annotations:
  • metaeuk/{assemblyid}/: MetaEuk protein and nucleotide fasta files of predicted genes, and GFF files
  • eggnog-mapper/{assemblyid}/: eggnog-mapper annotation files
  • clusters/: if using gene clustering, the fasta and cluster identities from MMSeqs clustering
• coverage/:
  • GOs/{sampleid}.GOSummary.csv: Per-sample abundances of each GO term in the supplied list
  • coverm/{sampleid}.txt: Per-gene read counts in TSV format
• GO_df_long.csv: GO summaries for all samples merged into one summary file
• taxonomy/{assemblyid}_taxdb/: MMseqs taxonomy DBs for the unfiltered assemblies.

Attribution

This pipeline uses code and infrastructure developed and maintained by the nf-core community, reused here under the MIT license.

The nf-core framework for community-curated bioinformatics pipelines.

Philip Ewels, Alexander Peltzer, Sven Fillinger, Harshil Patel, Johannes Alneberg, Andreas Wilm, Maxime Ulysse Garcia, Paolo Di Tommaso & Sven Nahnsen.

Nat Biotechnol. 2020 Feb 13. doi: 10.1038/s41587-020-0439-x. In addition, references of tools and data used in this pipeline are as follows: