Clinical and molecular epidemiology of carbapenemase-producing Enterobacterales in hospitalized patients

# Wiki

# Data location: 

The data was transferred from Athena medmicro):

``` 

/MedMicro/Clinton/CRE Pfizer Feb 2022/CRE study_1A_results_17022022 

/MedMicro/Clinton/CRE Pfizer Feb 2022/CRE study_1B_results_21022022

```

to Ilifu:

```

/scratch3/users/katiel/Clinton/CRE_study_August_2022/

```

# Reference data:

Klebsiella pneumoniae – strain HS11286 (GenBank accession no. CP003200.1) (n=18); 

Serratia marcescens – strain KS10 (GenBank accession no. CP027798.1) (n=3); 

Escherichia coli – strain ATCC 25922 (GenBank accession no. CP009072.1) (n=1); and 

Enterobacter cloacae – strain ATCC 13047 (GenBank accession no. NC_014121.1) (n=1).

```

/scratch3/users/katiel/Clinton/CRE_study_August_2022/ref_genomes

```

# Objectives workflow:

![workflow.png]()

# QC:

11 sample failed QC phred scores before trimming and filtering; none failed after filtering and trimming. Filtering and trimming were executed as follows:

```

nextflow run kviljoen/fastq_QC --reads '/scratch3/users/katiel/Clinton/CRE_study_August_2022/raw/study_1A_B_combined/*_R{1,2}_001.fastq.gz' -profile ilifu

```

QC reports can be found in the 'files' tab

# AMR profiling

The preference from Clinton is to do AMR profiling with the ResFinder DB. I'm getting errors there that I think relate to the header formatting though so in the interim have run with the ARG_annot DB that we used for previous projects as:

## ARGannot

```

nextflow run kviljoen/uct-srst2 --reads '/scratch3/users/katiel/Clinton/CRE_study_August_2022/2022-09-19-fastq_QC/bbduk/*_{1,2}.fq' -profile ilifu --gene_db /cbio/users/katie/Nicol/Ps_aerug_srst2_MLST/ARGannot_r3.fasta --outdir /scratch3/users/katiel/Clinton/CRE_study_August_2022/srst2_ARGannot/coverage_80_run --min_gene_cov 80

```

Individual results files compiled as:

```

srst2 --prev_output *results.txt --output ARGannot_AMRs

```

## CARD DB: 

This database is the recommended by srst2 and has been formatted by them already. The DB was downloaded with:

```

wget https://github.com/katholt/srst2/blob/master/data/CARD_v3.0.8_SRST2.fasta?raw=true -O CARD_v3.0.8_SRST2.fasta

```

Pipeline execution as:

```

nextflow run kviljoen/uct-srst2 --reads '/scratch3/users/katiel/Clinton/CRE_study_August_2022/2022-09-19-fastq_QC/bbduk/*_{1,2}.fq' -profile ilifu --gene_db /scratch3/users/katiel/Clinton/CRE_study_August_2022/ref_files/CARD_v3.0.8_SRST2.fasta --outdir /scratch3/users/katiel/Clinton/CRE_study_August_2022/srst2_CARD/coverage_80_run --min_gene_cov 80

```

Individual results files compiled as:

```

srst2 --prev_output *results.txt --output CARD_AMRs

```

# Virulence factors

Building the relevant VFDB for Klebsiella requires a python script that needs the biopython module (use the /cbio/users/katie/singularity_containers/srst2_v2.simg singularity container for this)

NB: in order to use the correct python version (2.7.5) for srst2 I first had to comment out the lines at the end of my .bashrc file relating to conda initialize

Build genus-specific DB:

```

python /cbio/users/katie/Nicol/Ps_aerug_srst2_MLST/VFDB/srst2/database_clustering/VFDBgenus.py --infile /scratch3/users/katiel/Clinton/CRE_study_August_2022/ref_files/VFDB_setB_nt.fas --genus Klebsiella 

```

was used to create the VF DB Klebsiella.fsa 

The same procedure (as last year ;) was executed for Escherichia, Serratia and Enterobacter

cd-hit (needed to build vfdb as outlined here https://github.com/katholt/srst2#using-the-vfdb-virulence-factor-database-with-srst2) docker images was pulled from here https://quay.io/repository/biocontainers/cd-hit?tab=tags and converted to singularity image on BST server:

```

singularity exec /cbio/users/katie/singularity_containers/cd-hit.simg /bin/bash

```

then run CD-HIT to cluster the sequences for this genus, at 90% nucleotide identity:

```

 cd-hit -i Klebsiella.fsa -o Klebsiella_cdhit90 -c 0.9 > Klebsiella_cdhit90.stdout

```

Repeat for other .fsa DBs

NExt parse the cluster output and tabulate the results using the specific Virulence gene DB compatible script (use srst2_v2.simg again)

```

python /cbio/users/katie/Nicol/Ps_aerug_srst2_MLST/VFDB/srst2/database_clustering/VFDB_cdhit_to_csv_KLedit.py --cluster_file Klebsiella_cdhit90.clstr --infile Klebsiella.fsa --outfile Klebsiella_cdhit90.csv

```

Next convert the resulting csv table to a SRST2-compatible sequence database using:

```

python /cbio/users/katie/Nicol/Ps_aerug_srst2_MLST/VFDB/srst2/database_clustering/csv_to_gene_db.py -t Klebsiella_cdhit90.csv -o Klebsiella_VF_clustered.fasta -s 5

```

The actual VF typing can now be done using this clustered DB:

```

nextflow run kviljoen/uct-srst2 --reads '/scratch3/users/katiel/Clinton/CRE_study_August_2022/2022-09-19-fastq_QC/bbduk/*_{1,2}.fq' -profile ilifu --gene_db /scratch3/users/katiel/Clinton/CRE_study_August_2022/ref_files/Klebsiella_VF_clustered.fasta --outdir /scratch3/users/katiel/Clinton/CRE_study_August_2022/srst2_VFs/coverage_80_run --min_gene_cov 80

```