ReferenceSeeker: rapid determination of appropriate reference genomes

Contents

• Description
• Input & Output
• Installation
  • BioConda
  • GitHub
• Usage
• Examples
• Databases
  • RefSeq
  • Custom
• Dependencies
• Citation
• Feedback

Description

ReferenceSeeker determines closely related reference genomes following a scalable hierarchical approach combining an fast kmer profile-based database lookup of candidate reference genomes and subsequent computation of specific average nucleotide identity (ANI) values for the rapid determination of suitable reference genomes.

ReferenceSeeker computes kmer-based genome distances between a query genome and potential reference genome candidates via Mash (Ondov et al. 2016). For resulting candidates ReferenceSeeker subsequently computes (bidirectional) ANI values picking genomes meeting community standard thresholds by default (ANI >= 95 % & conserved DNA >= 69 %) (Goris, Konstantinos et al. 2007) ranked by the product of ANI and conserved DNA values to take into account both genome coverage and identity.

Custom databases can be built with local genomes. For further convenience, we provide pre-built databases with sequences from RefSeq (https://www.ncbi.nlm.nih.gov/refseq), GTDB and PLSDB copmrising the following taxa:

• bacteria
• archaea
• fungi
• protozoa
• viruses

as well as plasmids.

The reasoning for subsequent calculations of both ANI and conserved DNA values is that Mash distance values correlate well with ANI values for closely related genomes, however the same is not true for conserved DNA values. A kmer fingerprint-based comparison alone cannot distinguish if a kmer is missing due to a SNP, for instance or a lack of the kmer-comprising subsequence. As DNA conservation (next to DNA identity) is very important for many kinds of analyses, e.g. reference based SNP detections, ranking potential reference genomes based on a mash distance alone is often not sufficient in order to select the most appropriate reference genomes. If desired, ANI and conserved DNA values can be computed bidirectionally.

Input & Output

Input

Path to a taxon database and a draft or finished genome in (zipped) fasta format:

bashCopy
$ referenceseeker ~/bacteria GCF_000013425.1.fna

Output

Tab separated lines to STDOUT comprising the following columns:

Unidirectionally (query -> references):

• RefSeq Assembly ID
• Mash Distance
• ANI
• Conserved DNA
• NCBI Taxonomy ID
• Assembly Status
• Organism

bashCopy
#ID    Mash Distance    ANI    Con. DNA    Taxonomy ID    Assembly Status    Organism
GCF_000013425.1    0.00000    100.00    100.00    93061    complete    Staphylococcus aureus subsp. aureus NCTC 8325
GCF_001900185.1    0.00002    100.00    99.89     46170    complete    Staphylococcus aureus subsp. aureus HG001
GCF_900475245.1    0.00004    100.00    99.57     93061    complete    Staphylococcus aureus subsp. aureus NCTC 8325 NCTC8325
GCF_001018725.2    0.00016    100.00    99.28     1280     complete    Staphylococcus aureus FDAARGOS_10
GCF_003595465.1    0.00185    99.86     96.81     1280     complete    Staphylococcus aureus USA300-SUR6
GCF_003595385.1    0.00180    99.87     96.80     1280     complete    Staphylococcus aureus USA300-SUR2
GCF_003595365.1    0.00180    99.87     96.80     1280     complete    Staphylococcus aureus USA300-SUR1
GCF_001956815.1    0.00180    99.87     96.80     46170    complete    Staphylococcus aureus subsp. aureus USA300_SUR1
...

Bidirectionally (query -> references [QR] & references -> query [RQ]):

• RefSeq Assembly ID
• Mash Distance
• QR ANI
• QR Conserved DNA
• RQ ANI
• RQ Conserved DNA
• NCBI Taxonomy ID
• Assembly Status
• Organism

bashCopy
#ID    Mash Distance    QR ANI    QR Con. DNA    RQ ANI    RQ Con. DNA    Taxonomy ID    Assembly Status    Organism
GCF_000013425.1    0.00000    100.00    100.00    100.00    100.00    93061    complete    Staphylococcus aureus subsp. aureus NCTC 8325
GCF_001900185.1    0.00002    100.00    99.89     100.00    99.89     46170    complete    Staphylococcus aureus subsp. aureus HG001
GCF_900475245.1    0.00004    100.00    99.57     99.99     99.67     93061    complete    Staphylococcus aureus subsp. aureus NCTC 8325 NCTC8325
GCF_001018725.2    0.00016    100.00    99.28     99.95     98.88     1280     complete    Staphylococcus aureus FDAARGOS_10
GCF_001018915.2    0.00056    99.99     96.35     99.98     99.55     1280     complete    Staphylococcus aureus NRS133
GCF_001019415.2    0.00081    99.99     94.47     99.98     99.36     1280     complete    Staphylococcus aureus NRS146
GCF_001018735.2    0.00096    100.00    94.76     99.98     98.58     1280     complete    Staphylococcus aureus NRS137
GCF_003354885.1    0.00103    99.93     96.63     99.93     96.66     1280     complete    Staphylococcus aureus 164
...

Installation

ReferenceSeeker can be installed via Conda and Git(Hub). In either case, a taxon database must be downloaded which we provide for download at Zenodo:
For more information have a look at Databases.

BioConda

The preferred way to install and run ReferenceSeeker is Conda using the Bioconda channel:

bashCopy
$ conda install -c bioconda referenceseeker
$ referenceseeker --help

GitHub

Alternatively, you can use this raw GitHub repository:

1. install necessary Python dependencies (if necessary)
2. clone the latest version of the repository
3. install necessary 3rd party executables (Mash, MUMmer4)

bashCopy
$ pip3 install --user biopython xopen
$ git clone https://github.com/oschwengers/referenceseeker.git
$ # install Mash & MUMmer
$ ./referenceseeker/bin/referenceseeker --help

Test

To test your installation we prepared a tiny mock database comprising 4 Salmonella spp genomes and a query assembly (SRA: SRR498276) in the tests directory:

bashCopy
$ git clone https://github.com/oschwengers/referenceseeker.git

  # GitHub installation
$ ./referenceseeker/bin/referenceseeker referenceseeker/test/db referenceseeker/test/data/Salmonella_enterica_CFSAN000189.fasta

  # BioConda installation
$ referenceseeker referenceseeker/test/db referenceseeker/test/data/Salmonella_enterica_CFSAN000189.fasta

Expected output:

bashCopy
#ID    Mash Distance    ANI    Con. DNA    Taxonomy ID    Assembly Status    Organism
GCF_000439415.1    0.00003    100.00    99.55    1173427    complete    Salmonella enterica subsp. enterica serovar Bareilly str. CFSAN000189
GCF_900205275.1    0.01522    98.61     83.13    90370      complete    Salmonella enterica subsp. enterica serovar Typhi

Usage

Usage:

bashCopy
usage: referenceseeker [--crg CRG] [--ani ANI] [--conserved-dna CONSERVED_DNA]
                       [--unfiltered] [--bidirectional] [--help] [--version]
                       [--verbose] [--threads THREADS]
                       <database> <genome>

Rapid determination of appropriate reference genomes.

positional arguments:
  <database>            ReferenceSeeker database path
  <genome>              target draft genome in fasta format

Filter options / thresholds:
  These options control the filtering and alignment workflow.

  --crg CRG, -r CRG     Max number of candidate reference genomes to pass kmer
                        prefilter (default = 100)
  --ani ANI, -a ANI     ANI threshold (default = 0.95)
  --conserved-dna CONSERVED_DNA, -c CONSERVED_DNA
                        Conserved DNA threshold (default = 0.69)
  --unfiltered, -u      Set kmer prefilter to extremely conservative values
                        and skip species level ANI cutoffs (ANI >= 0.95 and
                        conserved DNA >= 0.69
  --bidirectional, -b   Compute bidirectional ANI/conserved DNA values
                        (default = False)

Runtime & auxiliary options:
  --help, -h            Show this help message and exit
  --version, -V         show program's version number and exit
  --verbose, -v         Print verbose information
  --threads THREADS, -t THREADS
                        Number of used threads (default = number of available
                        CPU cores)

Examples

Installation:

bashCopy
$ conda install -c bioconda referenceseeker
$ wget https://zenodo.org/record/4415843/files/bacteria-refseq.tar.gz
$ tar -xzf bacteria-refseq.tar.gz
$ rm bacteria-refseq.tar.gz

Simple:

bashCopy
$ # referenceseeker <REFERENCE_SEEKER_DB> <GENOME>
$ referenceseeker bacteria-refseq/ genome.fasta

Expert: verbose output and increased output of candidate reference genomes using a defined number of threads:

bashCopy
$ # referenceseeker --crg 500 --verbose --threads 8 <REFERENCE_SEEKER_DB> <GENOME>
$ referenceseeker --crg 500 --verbose --threads 8 bacteria-refseq/ genome.fasta

Databases

ReferenceSeeker depends on databases comprising taxonomic genome informations as well as kmer hash profiles for each entry.

Pre-built

We provide pre-built databases based on public genome data hosted at Zenodo: :

RefSeq

release: 221 (2024-01-09)

Taxon	URL	# Genomes	Size
bacteria	https://zenodo.org/record/4415843/files/bacteria-refseq.tar.gz	50,226	59.6 Gb
archaea	https://zenodo.org/record/4415843/files/archaea-refseq.tar.gz	905	897 Mb
fungi	https://zenodo.org/record/4415843/files/fungi-refseq.tar.gz	557	5.9 Gb
protozoa	https://zenodo.org/record/4415843/files/protozoa-refseq.tar.gz	90	1.1 Gb
viruses	https://zenodo.org/record/4415843/files/viral-refseq.tar.gz	14,012	1 Mb

GTDB

release: v214 (2024-01-11)

Taxon	URL	# Genomes	Size
bacteria	n.a. due to storage quota resitrctions	80,789	82 Gb
archaea	https://zenodo.org/record/4415843/files/archaea-gtdb.tar.gz	4,416	2.8 Gb

Plasmids

In addition to the genome based databases, we provide the following plasmid databases based on RefSeq and PLSDB:

DB	URL	# Plasmids	Size
RefSeq	https://zenodo.org/record/4415843/files/plasmids-refseq.tar.gz	81,674	2.6 Gb
PLSDB	https://zenodo.org/record/4415843/files/plasmids-plsdb.tar.gz	59,882	2.3 Gb

Custom database

If above mentiond RefSeq based databases do not contain sufficiently-close related genomes or are just too large, ReferenceSeeker provides auxiliary commands in order to either create databases from scratch or to expand existing ones. Therefore, a second executable referenceseeker_db accepts init and import subcommands:

Usage:

bashCopy
usage: referenceseeker_db [--help] [--version] {init,import} ...

Rapid determination of appropriate reference genomes.

positional arguments:
  {init,import}  sub-command help
    init         Initialize a new database
    import       Add a new genome to database

Runtime & auxiliary options:
  --help, -h     Show this help message and exit
  --version, -V  show program's version number and exit

If a new database should be created, use referenceseeker_db init:

bashCopy
usage: referenceseeker_db init [-h] [--output OUTPUT] --db DB

optional arguments:
  -h, --help            show this help message and exit
  --output OUTPUT, -o OUTPUT
                        output directory (default = current working directory)
  --db DB, -d DB        Name of the new ReferenceSeeker database

This new database or an existing one can be used to import genomes in Fasta, GenBank or EMBL format:

bashCopy
usage: referenceseeker_db import [-h] --db DB --genome GENOME [--id ID]
                                 [--taxonomy TAXONOMY]
                                 [--status {complete,chromosome,scaffold,contig}]
                                 [--organism ORGANISM]

optional arguments:
  -h, --help            show this help message and exit
  --db DB, -d DB        ReferenceSeeker database path
  --genome GENOME, -g GENOME
                        Genome path [Fasta, GenBank, EMBL]
  --id ID, -i ID        Unique genome identifier (default sequence id of first
                        record)
  --taxonomy TAXONOMY, -t TAXONOMY
                        Taxonomy ID (default = 12908 [unclassified sequences])
  --status {complete,chromosome,scaffold,contig}, -s {complete,chromosome,scaffold,contig}
                        Assembly level (default = contig)
  --organism ORGANISM, -o ORGANISM
                        Organism name (default = "NA")

Example:

If ReferenceSeeker is properly installed, clone this repository and change into its parent directoriy.

$ git clone https://github.com/oschwengers/referenceseeker.git
$ cd referenceseeker
$ referenceseeker_db init --db test-db --output ./
$ referenceseeker_db import --db ./test-db --genome test/db/GCF_000439415.1.fna.gz --id GCF_000439415.1 --taxonomy 28901 --status complete --organism "Salmonella enterica subsp. enterica serovar Bareilly str. CFSAN000189"
$ referenceseeker_db import --db ./test-db --genome test/db/GCF_002211925.1.fna.gz --id GCF_002211925.1 --organism "Salmonella bongori str. SA19983605"
$ referenceseeker -v ./test-db ./test/data/Salmonella_enterica_CFSAN000189.fasta

Dependencies

ReferenceSeeker needs the following dependencies:

• Python (3.8, 3.9), Biopython (>=1.78), xopen(>=1.1.0)
• Mash (2.3) https://github.com/marbl/Mash
• MUMmer (4.0.0-beta2) https://github.com/gmarcais/mummer

ReferenceSeeker has been tested against aforementioned versions.

Citation

Schwengers et al., (2020). ReferenceSeeker: rapid determination of appropriate reference genomes. Journal of Open Source Software, 5(46), 1994, https://doi.org/10.21105/joss.01994

Feedback

We highly wellcome and appreciate feedback of all kind!

So, if you run into any issues with ReferenceSeeker, we'd be happy to hear about it! Please, start the pipeline with -v (verbose) and do not hesitate to file an issue here on GitHub including as much of the following as possible:

• a detailed description of the issue
• the ReferenceSeeker cmd line output
• a reproducible example of the issue with a small dataset that you can share (helps us identify whether the issue is specific to a particular computer, operating system, and/or dataset).

The maintenance of ReferenceSeeker is supported by deNBI. If you would like to provide (non-technical) feedback, please find a service monitoring survey here.