De-novo OTUs from shotgun metagenomes
Latest commit 39625c7 Feb 15, 2017 @wwood committed on GitHub Detail Docker installation method.
Failed to load latest commit information.

Welcome to SingleM.

SingleM is a tool to find the abundances of discrete operational taxonomic units (OTUs) directly from shotgun metagenome data, without heavy reliance of reference sequence databases. It is able to differentiate closely related species even if those species are from lineages new to science.

Where GraftM can give a taxonomic overview of your community e.g. proportion of a community from a particular taxonomic family, SingleM finds sequence-based OTUs from raw, untrimmed metagenomic reads.

This gives you the ability to answer questions such as:

  • How many different kinds of TM6 do I have?
  • What is the Chao 1 diversity of my sample?
  • Are the Acidobacteria in sample 1 very closely related to the Acidobacteria in sample 2?
  • Do I have population genomes for the main community members?
  • How diverse are the Pelagibacteria relative to the Flavobacteria?
  • Has my genome been observed in any samples submitted to the SRA?

Generating an OTU table

An overview of your community can be obtained like so. Please use raw metagenome reads, not QC'd reads. QC'ing reads often makes them too short for SingleM to use.

singlem pipe --sequences my_sequences.fastq.gz --otu_table otu_table.csv --threads 24

The output table consists of columns:

gene    sample  sequence        num_hits        coverage        taxonomy
4.21.ribosomal_protein_S19_rpsS my_sequences  TGGTCGCGCCGTTCGACGGTCACTCCGGACTTCATCGGCCTACAGTTCGCCGTGCACATC    1       1.64    Root; d__Bacteria; p__Proteobacteria; c__Deltaproteobacteria; o__Desulfuromonadales
4.21.ribosomal_protein_S19_rpsS my_sequences  TGGTCGCGGCGCTCAACCATTCTGCCCGAGTTCGTCGGCCACACCGTGGCCGTTCACAAC    1       1.64    Root; d__Bacteria; p__Acidobacteria; c__Solibacteres; o__Solibacterales; f__Solibacteraceae; g__Candidatus_Solibacter; s__Candidatus_Solibacter_usitatus
  1. marker name
  2. sample name
  3. sequence of the OTU
  4. number of reads detected from that OTU
  5. estimated coverage of a genome from this OTU
  6. "median" taxonomic classification of each of the reads in the OTU according to pplacer

Currently SingleM concentrates on 15 single copy marker genes to provide fine-grained differentiation of species that is independent of the copy-number variation issues that hamper 16S analyses. SingleM is reasonably fast and is quite scalable, although there is much room for improvement. On average, each of the 15 genes better differentiates closely related lineages than a typical 16S amplicon-based study.

Further processing of OTU tables

Summarising OTU tables

Once an OTU table has been generated with the pipe command, it can be further processed in various ways using summarise:

Create a Krona plot of the community. The following command generates my_krona*.html files which can be viewed in a web browser:

singlem summarise --input_otu_table otu_table.csv --krona my_krona

Cluster sequences, collapsing them into OTUs with less resolution, but with more robustness against sequencing error:

singlem summarise --input_otu_table otu_table.csv --cluster --clustered_output_otu_table clustered.otu_table.csv

Rarefy a set of OTU tables so that each sample contains the same number of OTU sequences:

singlem summarise --input_otu_tables otu_table.csv other_samples.otu_table.csv --rarefied_output_otu_table rarefied.otu_table.csv --number_to_choose 100

Calculating beta diversity between samples

As SingleM generates OTUs that are independent of taxonomy, they can be used as input to beta diversity methods known to be appropriate for the analysis of 16S amplicon studies, of which there are many. We recommend express beta diversity (EBD) as it implements many different metrics with a unified interface. For instance to calculate Bray-Curtis beta diversity, first convert your OTU table to unifrac format using singlem summarise:

singlem summarise --input_otu_table otu_table.csv --unifrac otu_table.unifrac

The above commands generates 15 different unifrac format files, one for each marker gene used in SingleM. At this point, you need to choose one table to proceed with. Hopefully, the choice matters little, but it might pay to use multiple tables and ensure that the results are consistent.

To calculate beta diversity, use the EBD script to convert the unifrac format into ebd format, and calculate the diversity metric: otu_table.unifrac.4.12.ribosomal_protein_L11_rplK.unifrac otu_table.ebd
ExpressBetaDiversity -s otu_table.ebd -c Bray-Curtis

Phylogenetic tree-based methods of calculating beta diversity can also be calculated, but pipe must be used to generate a new OTU table using the diamond_example taxonomy assignment method so that each OTU is assigned to a single leaf in the tree:

singlem pipe --sequences my_sequences.fastq.gz --otu_table otu_table.diamond_example.csv --threads 24 --assignment_method diamond_example
singlem summarise --otu_tables otu_table.diamond_example.csv --unifrac otu_table.diamond_example.csv otu_table.diamond_example.unifrac otu_table.diamond_example.ebd
ExpressBetaDiversity -s otu_table.diamond_example.ebd -c Bray-Curtis -t `singlem get_tree --marker_name 4.21.ribosomal_protein_S19_rpsS`

Creating and querying SingleM databases

It can be useful in some situations to search for sequences in OTU tables. For instance, you may ask "is the most abundant OTU or anything similar in samples B, C or D?" To answer this question make a SingleM database from sample B, C & D's OTU tables:

singlem makedb --otu_tables sample_B.csv sample_C.csv sample_D.csv --db_path sample_BCD.sdb

.sdb is the conventional file extension for SingleM databases. Then to query this database


Appraising genome recovery efforts

To assess how well a set of genomes (or population genomes) represent those present in a metagenome, first run pipe on both the genomes and the raw reads, and then use appraise:

singlem appraise --metagenome_otu_tables metagenome.otu_table.csv --genome_otu_tables genomes.otu_table.csv

One may also accommodate some sequence differences, with --imperfect, or output OTU tables of OTUs in the genomes or not in the genomes with --accounted_for_otu_table and --unaccounted_for_otu_table.


Installation via GNU Guix

The most straightforward way of installing SingleM is to use the GNU Guix package which is part of the ACE Guix package collection. This method installs not just the Python libraries required but the compiled bioinformatics tools needed as well. Once you have installed Guix, clone the ACE collection and install:

git clone
GUIX_PACKAGE_PATH=ace-guix guix package --install singlem

Beyond installing GNU Guix, super-user privileges are not required.

Installation via DockerHub

A docker image generated from the Guix package is available on DockerHub. After installing Docker:

docker pull wwood/singlem

If the sequence data to be analyzed is in the current working directory, SingleM can be used like so:

docker run -v `pwd`:`pwd` wwood/singlem singlem pipe --sequences `pwd`/my.fastq.gz --otu_table `pwd`/my.otu_table.csv --threads 15

Installation via PyPI

To install the Python libraries required:

pip install graftm
pip install singlem

You may need super-user privileges.

SingleM also has the following non-Python dependencies:

Some dependencies of GraftM:


If you have any questions or comments, send a message to the SupportM mailing list or raise a GitHib issue.


SingleM is written by Ben Woodcroft (@wwood) at the Australian Centre for Ecogenomics (UQ) and is licensed under GPL3 or later.