emmtyper is a command line tool for emm-typing of Streptococcus pyogenes using a de novo or complete assembly.
By default, we use the U.S. Centers for Disease Control and Prevention trimmed emm subtype database, which can be found here. The database is curated by Dr. Velusamy Srinivasan. We take this opportunity to thank Dr. Srinivasan for his work.
The tool has two basic modes:
blast: In this mode the contigs are blasted against the trimmed FASTA database curated by the CDC.pcr: In this mode, first an in silico PCR is done on the contigs using theisPCRtool (Kuhn et al. 2013). The resulting fragments are then blasted against the trimmed FASTA database curated by the CDC. Two sets of primers are provided for the user to choose from: 1. The canonical CDC primers used for conventional PCR (Whatmore and Kehoe 1994); 2. the primers described by Frost et al. 2020. This last set uses the forward primers of Whatmore and Kehoe (1994), but provide a re-designed reverse primer. There is also the option of the user providing their own primer set in the format required by theisPCRtool.
The difficulty in performing M-typing is that there is a single gene of interest (emm), but two other homologue genes (enn and mrp), often referred to as emm-like. The homologue genes may or may not occur in the isolate of interest. When performing emm-typing from an assembly, we can distinguish betweeen one or more clusters of matches on the contigs. The best match for each of the clusters identified is then parsed from the BLAST results. Where possible, we try to distinguish between matches to the emm gene, and matches to one of the emm-like genes.
Possible arrangments:
emm
---->>>>>>>----
mrp emm enn
---->>>>>>----->>>>>>------>>>>>>-----
emm enn
----->>>>>>------>>>>>>-----
blastn≥ 2.6 (tested on 2.9)isPcrpython≥ 3.6
brew install python blast ispcr
pip3 install emmtyper
emmtyper --helpconda install -c conda-forge -c bioconda -c defaults emmtyperemmtyper has two workflows: directly BLASTing the contigs against the DB, or using isPcr to generate an in silico PCR product that is then BLASTed against the DB. The BLAST results go through emmtyper's business logic to distinguish between emm and emm-like alleles and derive the isoolate M-type.
The basic usage of emmtyper is in the form of:
emmtyper [options] contig1 contig2 ... contigNAll the available options can be printed to the console with emmtyper --help. Options passed on to blast are tagged with [BLAST], and those for isPcr are tagged with [isPcr].
Usage: emmtyper [OPTIONS] [FASTA]...
Welcome to emmtyper.
Usage:
emmtyper *.fasta
Options:
--version Show the version and exit.
-w, --workflow [blast|pcr] Choose workflow [default: blast]
-d, --blast_db TEXT Path to EMM BLAST DB [default:
/path/to/emmtyper/db/emm.fna]
-k, --keep Keep BLAST and isPcr output files.
[default: False]
-d, --cluster-distance INTEGER Distance between cluster of matches to
consider as different clusters. [default:
500]
-o, --output TEXT Output stream. Path to file for output to a
file. [default: stdout]
-f, --output-format [short|verbose|visual]
Output format.
--dust [yes|no|level window linker]
[BLAST] Filter query sequence with DUST.
[default: no]
--percent-identity INTEGER [BLAST] Minimal percent identity of
sequence. [default: 95]
--culling-limit INTEGER [BLAST] Total hits to return in a position.
[default: 5]
--mismatch INTEGER [BLAST] Threshold for number of mismatch to
allow in BLAST hit. [default: 4]
--align-diff INTEGER [BLAST] Threshold for difference between
alignment length and subject length in BLAST
hit. [default: 5]
--gap INTEGER [BLAST] Threshold gap to allow in BLAST hit.
[default: 2]
--blast-path TEXT [BLAST] Specify full path to blastn
executable.
--pcr-primers [cdc|frost] [isPcr] Primer set to use (either canonical
CDC or Frost et al. 2020). [default: cdc]
--primer-db TEXT [isPcr] PCR primer. Text file with 3
columns: Name, Forward Primer, Reverse
Primer. This options overrides --pcr-
primers.
--min-perfect INTEGER [isPcr] Minimum size of perfect match at 3\'
primer end. [default: 15]
--min-good INTEGER [isPcr] Minimum size where there must be 2
matches for each mismatch. [default: 15]
--max-size INTEGER [isPcr] Maximum size of PCR product.
[default: 2000]
--ispcr-path TEXT [isPcr] Specify full path to isPcr
executable.
--help Show this message and exit.Most of these options are self explanatory. The two expections are:
--workflow: choose between ablastonly workflow, or a in silico PCR followed byblastworkflow. See below for more information.--clust_distancedefines the minimum distance between clusters of matched sequences on the contigs to generate separateemm-typecalls for each clusters. Clusters of matches that are within the minimumclust-distanceare treated as a single location match.--output_typedemonstrated below.
# basic call using the blast workflow for a single contig file
emmtyper isolate1.fa
# basic call using the pcr workflow for all the .fa files in a folder
emmtyper -w pcr *.fa
# basic call changing some of the options for blast
emmtyper --keep --culling_limit 10 --align_diff 10 *.fa
# call using the pcr workflow changing some of the isPcr options and
# using the visual output format - this will run the with the CDC canonical
# primers
emmtyper -w pcr --output-format visual --max-size 2000 --mismatch 5 *.fa
# same as above, but now with the Frost et al. 2020 primers.
emmtyper -w pcr --output-format visual --max-size 2000 --mismatch 5 --pcr-primers frost *.faemmtyper has three different result formats: short, verbose, and visual.
emmtyper by default produces the short version. This consists of five values in tab-separated format printed to stdout.
The values are:
- Isolate name
- Number of clusters: should be between 1 and 3, larger values could indicate contamination
- Predicted
emm-type - Possible
emm-likealleles (semi-colon separated list) - EMM cluster: Functional grouping of EMM types into 48 clusters
The verbose result returns:
- Isolate name
- Number of BLAST hits
- Number of clusters: should be between 1 and 3, larger values could indicate contamination
- Predicted
emm-type - Position(s)
emm-likealleles in the assembly - Possible
emm-likealleles (semi-colon separated list) emm-likeposition(s) in assembly- EMM cluster: Functional grouping of EMM types into 48 clusters
The positions in the assembly are presented in the following format <contig_number>:<position_in_contig>.
The visual result returns an ASCII map of the emm and, if found any emm-alleles, in the genome. Alleles on a single contig are separated by "-", with each "-" representing 500bp. Alleles found on different contigs are separated with tab.
The alleles can be tagged with a suffix character to indicate different possibilities:
| Tag | Description | Additional Information |
|---|---|---|
| * | Suspect emm-like |
Allele flagged in the CDC database as possibly emm-like |
| ~ | Imperfect score | Match score below 100% |
Example for all result format:
Short format:
Isolate1 1 EMM65.0 E6
Isolate2 3 EMM4.0 EMM236.3*;EMM156.0* E1
Isolate3 2 EMM52.1 EMM134.2* D4
Verbose format:
Isolate1 6 1 EMM65.0 5:82168 E6
Isolate2 8 3 EMM4.0 2:104111 EMM236.3*;EMM156.0* 2:102762;2:105504 E1
Isolate3 5 2 EMM52.1 14:10502 EMM134.2* 5:913 D4
Visual format:
Isolate1 EMM65.0
Isolate2 EMM156.0*--EMM4.0--EMM236.3*
Isolate3 EMM52.1 EMM134.2*
If you are not sure which pipeline to choose from, we recommend using blast first. The blast workflow is fast and works well with assemblies. You can then use the pcr mode if you wish to perform some troubleshooting.
For example, the pcr workflow might be useful when troubleshooting isolates for which emmtyper has reported more than 3 clusteres and/or too many alleles.
An important thing to note is that not all emm-like alleles can be identified by using by PCR typing. The pcr workflow can be used to test which hits would be returned if carrying out conventional M-typing using PCR. However, the workflow is not foolproof, as in silico PCR will fail when one or both primers do not align in the same contig (i.e., the allele is broken across two or more contigs) or there are mutations in the primer sites. In the former case, this might be an indication of poor sequence coverage or contamination.
We compared emmtyper against Sanger sequencing data and PHE's tool emm-typing-tool.
You can check out the validation comparison go to out binder:
- Andre Tan
- Torsten Seemann
- Jake Lacey
- Mark Davies
- Liam Mcintyre
- Hannah Frost
- Deborah Williamson
- Anders Gonçalves da Silva
The codebase for emmtyper was primarly written by Andre Tan as part of his Master's
Degree in Bioinformatics. Torsten Seemann, Deborah Williamson, and Anders Gonçalves da Silva provided supervision and assistance.
Hannah Frost contributed with EMM clustering by suggesting we incorporate it in to the code, and providing the necessary information to do so and test it.
Jake Lacey, Liam Mcintyre, and Mark Davies provided assistance in validating emmtyper.
The code is actively maintained by MDU Bioinformatics Team.
Contact the principal maintainer at andersgs at gmail dot com.
Please post bug reports, questions, suggestions in the Issues section.
Frost, H. R., Davies, M. R., Velusamy, S., Delforge, V., Erhart, A., Darboe, S., Steer, A., Walker, M. J., Beall, B., Botteaux, A., & Smeesters, P. R. (2020). Updated emm-typing protocol for Streptococcus pyogenes. Clinical Microbiology and Infection: The Official Publication of the European Society of Clinical Microbiology and Infectious Diseases, 26(7), 946.e5–e946.e8.
Kuhn, R. M., Haussler, D., & Kent, W. J. (2013). The UCSC genome browser and associated tools. Briefings in Bioinformatics, 14(2), 144–161.
Whatmore, A. M., & Kehoe, M. A. (1994). Horizontal gene transfer in the evolution of group A streptococcal emm-like genes: gene mosaics and variation in Vir regulons. Molecular Microbiology, 11(2), 363–374.