KleTy is a tool to type Klebsiella genome assemblies for:
- core genome MLST (cgMLST) for detailed genotyping of the core genome
- Hierarchical clusters (HierCC) that represents natural population
- Plasmid prediction and classification (PC)
- hypervirulence associated loci
- antimicrobial resistance determinants
KleTy: integrated typing scheme for core genome and plasmids reveals repeated emergence of multi-drug resistant epidemic lineages in Klebsiella worldwide Heng Li, Xiao Liu, Shengkai Li, Jie Rong, Shichang Xie, Yuan Gao, Ling Zhong, Quangui Jiang, Guilai Jiang, Yi Ren, Wanping Sun, Yuzhi Hong, Zhemin Zhou medRxiv 2024.04.16.24305880; doi: https://doi.org/10.1101/2024.04.16.24305880
KleTy was developed and tested in Python >=3.8. It depends on several Python libraries:
click
numba
numpy
pandas
biopython
pyarrow
fastparquet
All libraries can be installed using pip:
pip install click numba numpy pandas biopython pyarrow fastparquet
KleTy also calls NCBI-BLAST+:
ncbi-blast+
Which can be installed via 'apt' in UBUNTU:
sudo apt install -y ncbi-blast+
The whole environment can also be installed in conda:
conda create --name dty python==3.11
conda activate dty
conda install -c conda-forge biopython numba numpy pandas click pyarrow fastparquet
conda install -c bio-conda blast
The installation process normally finishes in <10 minutes.
NOTE: Please make sure that "makeblastdb" and "blastn" are all in the PATH environment variable (can be run without pointing to their actual location).
When run for the first time, KleTy will automatically download the reference plasmids from https://zenodo.org/records/12590507/files/plasmids.repr.fas.gz This will only run once. But note that the file is fairly large (816 MB), and will take a long time to download.
Alternatively, for those who have difficulty downloading the file within the pipeline. Please download the file by yourself, and copy it into "db/" under the KleTy folder. Then run
gzip -d plasmids.repr.fas.gz
makeblastdb -in plasmids.repr.fas -dbtype nucl
to generate the required database.
$ cd /path/to/KleTy/
$ python KleTy.py -q examples/CP015990.fna
The whole calculation finishes in ~1 minutes with 8 CPU threads (~2.5 minutes with one CPU thread). The screen output will be like:
07/04/2024 04:44:56 AM Running query: examples/CP015990.fna
07/04/2024 04:44:56 AM Searching VF/STRESS genes...
07/04/2024 04:45:06 AM Done.
07/04/2024 04:45:06 AM Searching AMR genes...
07/04/2024 04:45:16 AM Done.
07/04/2024 04:45:16 AM Searching plasmids...
07/04/2024 04:45:51 AM Done.
07/04/2024 04:45:51 AM Running cgMLST...
07/04/2024 04:46:09 AM Done.
And there are two outputs (see below for explanation):
CP015990.KleTy
CP015990.cgMLST.profile.gz
Usage: KleTy.py [OPTIONS]
Options:
-q, --query TEXT query genome in fasta or fastq format. May be
gzipped.
--ql TEXT a list of query files. One query per line.
-o, --prefix TEXT prefix for output. Only work when there is only one
query. default: query filename
-n, --n_proc INTEGER number of process to use. default: 8
-f, --plasmid_fragment flag to predict plasmid fragment sharing < 50% with
the reference
-m, --skip_gene flag to skip AMR/VF searching. default: False
-g, --skip_cgmlst flag to skip cgMLST. default: False
-p, --skip_plasmid flag to skip plasmid typing. default: False
--help Show this message and exit.
| Parameter | Explanation |
|---|---|
| -q, --query | Query genome. This can be in Fasta or Fastq format, and can be in plain text or GZIPped. |
| --ql | A list of query files. One query genome (file location) per line. KleTy will run these queries one by one and concatenate the outputs together. |
| -o, --prefix | Prefix for the outputs. There will be two files .KleTy and .cgMLST.profile.gz. Will use the prefix of the query file (or the ql file) if not specified. |
| -n, --n_proc | Number of processes to use. Default: 8 |
| -f, --plasmid_fragment | Flag to predict less reliable plasmid fragments that share <50% (but >=30%) of the reference plasmid. |
| -m, --skip_gene | Flag to skip AMR/VF Searching. This step normaly taks ~ 15 seconds. |
| -g, --skip_cgmlst | Flag to skip cgMLST calling. This step normaly taks ~ 20 seconds. |
| -p, --skip_plasmid | Flag to skip plasmid prediction. This step normaly taks ~ 30 seconds. |
<prefix>.KleTy
$ cat CP015990.KleTy
INPUT REPLICON SPECIES HC1360.500.200.100.50.20.10.5.2 REFERENCE PLASTYPE COVERAGE AMR:AMINOGLYCOSIDE AMR:BETA-LACTAM AMR:CARBAPENEM AMR:ESBL AMR:INHIBITOR-RESISTANT AMR:COLISTIN AMR:FOSFOMYCIN AMR:MACROLIDE AMR:PHENICOL AMR:QUINOLONE AMR:RIFAMYCIN AMR:GLYCOPEPTIDES AMR:SULFONAMIDE AMR:TETRACYCLINE AMR:TIGECYCLINE AMR:TRIMETHOPRIM AMR:BLA_INTRINSIC STRESS:COPPER STRESS:MERCURY STRESS:NICKEL STRESS:SILVER STRESS:TELLURIUM STRESS:ARSENIC STRESS:FLUORIDE STRESS:QUATERNARY_AMMONIUM VIRULENCE:clb VIRULENCE:iro VIRULENCE:iuc VIRULENCE:rmp VIRULENCE:ybt Others REPLICON:INC_TYPE REPLICON:MOB_TYPE REPLICON:MPF_TYPE ANNOTATION CONTIGS
examples/CP015990.fna ALL Klebsiella_pneumoniae 10.10.10.10.ND.ND.ND.ND.ND KLE_DA0156AA_AS - - aac(3)-IId^,aac(6')-Ib-cr.v2^,aadA16* OXA-1 KPC-2 - - - - mphA catB3.v2 GyrA-83F,GyrA-87A,ParC-80I,qnrA3^ arr-3 - sul1 - - dfrA27 SHV-28^ - merA,merE,merR_Ps,merT - - - - - qacEdelta1 - - - - fyuA_26,irp1_275,irp2_30,ybtA_78,ybtE_58,ybtP_75,ybtQ_88,ybtS_115,ybtT_26,ybtU_129,ybtX_73 - IncR - MPF_T - -
examples/CP015990.fna P1 - - CP059309.1 PT_361,PC_361 84.9 aac(6')-Ib-cr.v2^,aadA16* OXA-1 KPC-2 - - - - mphA catB3.v2 qnrA3^ arr-3 - sul1 - - dfrA27 -- merA,merE,merR_Ps,merT - - - - - qacEdelta1 - - - - - - IncR - - Klebsiella_pneumoniae_strain_Kp46596_plasmid_pKp46596-3,_complete_sequence CP015991.1
examples/CP015990.fna Others - - - - - aac(3)-IId^ - KPC-2 - - - - mphA - GyrA-83F,GyrA-87A,ParC-80I - - - - - - SHV-28^ - -- - - - - - - - - - fyuA_26,irp1_275,irp2_30,ybtA_78,ybtE_58,ybtP_75,ybtQ_88,ybtS_115,ybtT_26,ybtU_129,ybtX_73 - - - MPF_T - -
The columns are:
| Column | Explanation |
|---|---|
| INPUT | Filename of the input. Used to recognize query assemblies |
| REPLICON | Type of the replicon. It can be: "ALL" - A summary of the query. "P" - One plasmid per row. "Others" - Summary of the AMR/VF genes that are not in plasmids (likely carried by the chromosome). |
| SPECIES | Species designation of the query, inferred based on its cgMLST profile. Will not be reported with '-g'. |
| HC1360.500.200.100.50.20.10.5.2 | HierCC cluster designation of the query based on the cgMLST profile. HC1360 approximately equals to clonal complex (CC) in MLST. Lower HC levels were used for sub-population clusterings. Numbers after HC indicate the criteria of the single-linkage clustering. Will not be reported with '-g'. |
| REFERENCE | Accession of the reference for predicted plasmid. Will not be reported with '-p'. |
| PLASTYPE | PT (plasmid type) and PC (plasmid cluster) of the predicted plasmid. Will not be reported with '-p'. |
| COVERAGE | Coverage of the plasmid to the reference. Will not be reported with '-p'. |
| AMR:AMINOGLYCOSIDE | Predicted genes/mutations encoding resistance to AMINOGLYCOSIDE. |
| AMR:BETA-LACTAM | Predicted genes/mutations encoding resistance to BETA-LACTAM. |
| AMR:CARBAPENEM | Predicted genes/mutations encoding resistance to CARBAPENEM. |
| AMR:ESBL | Predicted genes/mutations encoding Extended-spectrum beta-lactamases (ESBLs). |
| AMR:INHIBITOR-RESISTANT | Predicted genes/mutations encoding resistance to Beta-Lactamase inhibitors. |
| AMR:COLISTIN | Predicted genes/mutations encoding resistance to COLISTIN. |
| AMR:FOSFOMYCIN | Predicted genes/mutations encoding resistance to FOSFOMYCIN. |
| AMR:MACROLIDE | Predicted genes/mutations encoding resistance to MACROLIDE. |
| AMR:PHENICOL | Predicted genes/mutations encoding resistance to PHENICOL. |
| AMR:QUINOLONE | Predicted genes/mutations encoding resistance to QUINOLONE. |
| AMR:RIFAMYCIN | Predicted genes/mutations encoding resistance to RIFAMYCIN. |
| AMR:GLYCOPEPTIDES | Predicted genes/mutations encoding resistance to GLYCOPEPTIDES. |
| AMR:SULFONAMIDE | Predicted genes/mutations encoding resistance to SULFONAMIDE. |
| AMR:TETRACYCLINE | Predicted genes/mutations encoding resistance to TETRACYCLINE. |
| AMR:TIGECYCLINE | Predicted genes/mutations encoding resistance to TIGECYCLINE. |
| AMR:TRIMETHOPRIM | Predicted genes/mutations encoding resistance to TRIMETHOPRIM. |
| AMR:BLA_INTRINSIC | Predicted intrinsic beta-lactamase in Klebsiella. |
| STRESS:COPPER | Predicted genes encoding resistance to COPPER. |
| STRESS:MERCURY | Predicted genes encoding resistance to MERCURY. |
| STRESS:NICKEL | Predicted genes encoding resistance to NICKEL. |
| STRESS:SILVER | Predicted genes encoding resistance to SILVER. |
| STRESS:TELLURIUM | Predicted genes encoding resistance to TELLURIUM. |
| STRESS:ARSENIC | Predicted genes encoding resistance to ARSENIC. |
| STRESS:FLUORIDE | Predicted genes encoding resistance to FLUORIDE. |
| STRESS:QUATERNARY_AMMONIUM | Predicted genes encoding resistance to QUATERNARY_AMMONIUM. |
| VIRULENCE:clb | colibactin (clb) |
| VIRULENCE:iro | salmochelin (iro) |
| VIRULENCE:iuc | aerobactin (iuc) |
| VIRULENCE:rmp | hypermucoidy (rmpA, rmpA2) |
| VIRULENCE:ybt | yersiniabactin (ybt) |
| Others | Other resistances |
| REPLICON:INC_TYPE | INC type of the plasmid. |
| REPLICON:MOB_TYPE | MOB type of the plasmid. |
| REPLICON:MPF_TYPE | MPF type of the plasmid. |
| ANNOTATION | Annotations of the predicted plasmids. |
| CONTIGS | Contigs associated with the predicted plasmids. |
This file can be used as inputs for GrapeTree(https://achtman-lab.github.io/GrapeTree/MSTree_holder.html) when ungzipped.
All data required for reproduction of the analysis were distributed in this repository under https://github.com/zheminzhou/KleTy/tree/main/db
These includes:
- plasmids.repr.clu.gz - IMPORTANT. A mapping table that specifies correlations between plasmids and PT/PCs.
- HierCC.tsv.gz - A tab-delimited table consisting of HierCC results for all ~70,000 genomes
- klebsiella.cgmlst - A list of core genes used in the dcgMLST scheme
- klebsiella.refsets.fas.gz - reference alleles for all pan genes (for calling new alleles)
- klebsiella.species - A mapping table that specifies correlations between genomes and Klebsiella species
- profile.parq - Allelic profiles of all ~70,000 genomes in parquet format, and can be read using the Pandas library (https://pandas.pydata.org/docs/reference/api/pandas.read_parquet.html).
- stress_CDS.gz - reference sequences for resistance to metal/biocides
- traditional_lasmid_type.fas.gz - reference sequences for INC/MOB/MPF types of the plasmids.
- kleborate/* - reference sequences from kleborate.