3a. Running DRAM

DRAM annotate

Once DRAM is set up you are ready to annotate some MAGs. The following command will generate your full annotation:

DRAM.py annotate -i 'my_bins/*.fa' -o annotation

my_bins should be replaced with the path to a directory which contains all of your bins you would like to annotated. If you only need to annotated a single genome (or an entire assembly) a direct path to a nucleotide fasta should be provided. Using 20 processors DRAM.py takes about 17 hours to annotate ~80 MAGs of medium quality or higher from a mouse gut metagenome.

In the output annotation folder there will be various files. genes.faa and genes.fna are fasta files with all genes called by prodigal with additional header information gained from the annotation as nucleotide and amino acid records respectively. genes.gff is a GFF3 with the same annotation information as well as gene locations. scaffolds.fna is a collection of all scaffolds/contigs given as input to DRAM.py annotate with added bin information in the headers. annotations.tsv is the most important output of the annotation. This includes all annotation information about every gene from all MAGs. Each line is a different gene and each column contains annotation information. trnas.tsv contains a summary of the tRNAs and rrnas.tsv contains a summary of the rRNAs found in each MAG and.

DRAM annotate Parameters

• Input files
  • Input fasta: fasta file or a string with wildcards (e.g. ‘MAGs/*.fa’) that leads to multiple fastas
  • Output directory: Folder to be created that will store output files
• Parameters
  • Minimum contig size: 2500
  • Minimum bit score for MMSeqs2 searches: 60
  • Minimum bit score for reverse best hit MMSeqs2 searches: 350
  • Number of threads to use: 10

DRAM annotate Outputs

• Tab separated file (.tsv) with all the annotations from Pfam, KEGG, UniProt, dbCAN, and MEROPS databases for all genes in all the input genomes
• GenBank files for each genome
• Single gene-finding format (.gff) file of all annotations across genomes
• Single fasta format file (.fasta) of each open reading frame nucleotide sequence and best ranked annotation (see Annotation grades section)
• Single fasta format file (.fasta) of each translated open reading frame amino acid sequence and best ranked annotation KEGG annotation
• Tab separated files (.tsv) with tRNAs and rRNAs

DRAM distill

Then after your annotation is finished you can summarize these annotations with the following command:

DRAM.py distill -i annotation/annotations.tsv -o genome_summaries --trna_path annotation/trnas.tsv --rrna_path annotation/rrnas.tsv

This command will generate three files. The first is called genome_summary.xlsx this contains a summary of metabolisms present in each genome. It gives gene by gene information across various metabolisms for every genome in your dataset. The genome_statistics.tsv file contains all measures required by the MIMAG about each fasta used as input. Finally the liquor.html is an interactive html that allows users to hover over each box to see what genes prompted the box color (Example here) and was manually curated to consider alternate genes for pathways and single processes. This heatmap allows the user to quickly profile ecosystem relevant process across hundreds of genomes.

DRAM distill Parameters

• Input files
  • Annotations: annotations.tsv file generated during the annotate step
  • Output directory: Directory to create and write outputs
  • rRNA path: rrnas.tsv file generated during the annotation step (optional)
  • tRNA path: trnas.tsv file generated during the annotation step (optional)
• Default Parameters
  • Grouping: fasta column from annotations.tsv file

DRAM distill Outputs

• Distillate
  • Tab separated file (.tsv) with genome statistics for all input genomes including all statistics required by recently defined MIMAG standards (Bowers, et al.)
  • Tab separated file (.tsv) with metabolism summary of all input genomes, which gives gene counts of functional and structural genes across a wide variety of metabolisms
• Liquor
  • HTML (.html) file containing an interactive heatmap showing coverage of pathways, the coverage of electron transport chain components, and the presence of selected metabolic functions
  • Tab separated file (.tsv) with corresponding genes for the html heatmap.

DRAM strainer

After you have completed your annotation and distillation you may want to further analyze genes of interest by making trees or functional modeling. To pull the genes you can use DRAM.py strainer. For example if you want to pull all pmoa/amoa genes based on KEGG annotations so you can make a tree:

DRAM.py strainer --identifiers K10944 -i annotations.tsv -f genes.faa -o amoa_pmoa_genes.faa

Or you might want to blast a few specific genes:

DRAM.py strainer --genes bin.2_scaffold_2_3 bin.4_scaffold_12_42 -i annotations.tsv -f genes.fna -o my_genes.fna

Or maybe you only want to see genes that are involved in glycolysis or the TCA cycle that are from bins from the Roseburia genus:

DRAM.py strainer -i hmp_bins/annotations.tsv -f hmp_bins/genes.fna -o genes.roseburia.glycoloysis_tca.fna --taxonomy g__Roseburia --categories glycolysis TCA

DRAM strainer parameters

• Input files:
  • Annotations: annotations.tsv file generated during the annotate step
  • Input fasta: genes fasta file (.faa or .fna) to be filtered
  • Output fasta: location to save filtered fasta file
• Default Parameters:
  • Fastas: None
  • Scaffolds: None
  • Genes: None
  • Identifiers: None