EVALUACIÓN ECONÓMICA

Dried down DNA samples were sent to University of Technology, Sydney for sample preparation and whole genome sequencing using the standard Illumina protocol for Miseq which was performed by Dr Michael Liu. Briefly DNA was suspended in nuclease free water. DNA was then fragmented to produce 5’ and 3’ recessed, overhang and blunt ends and tagmented using the Illumina Nextera kit which

combines fragmenting DNA with tagmentation of adaptors in one step: 8 µl

tagmentation buffer, 5 µl gDNA and ddH2O, 4 µl Nextera tagmentation mix. The

reaction was stopped using 4µl NT buffer after holding at 55 o_{C for 5 minutes and}

room temperature for a further 10 minutes. Samples were incubated with NT buffer for 5 minutes at room temperature to ensure the tagementation step has stopped.

The library was enriched through PCR amplification using 20 µl of the tagmented

library, 22µl KAPA PCR master mix, 1µl index 1 (i7) primers, 1 µl index 2 (i5) primers

and then vortexing to mix. The cycle for tagemented library amplification is as

follows: 72 oC for 3 minutes, 95 oC for 30 seconds, 12 cycles of 95 oC for 10 seconds,

55 oC for 30 seconds and 72 oC for 30 seconds, then 72 oC for 5 minutes and infinite

hold at 10 oC. Tagmented and amplified DNA samples were run on the BioAnalyzer

to check sizing and quantification of libraries.

After confirmation of DNA libraries solid phase reversible immobilisation (SPRI) bead clean-up was performed on pooled samples to select for right and left side. Samples were once again run on the bioannalyser to check cleaned sample. DNA library was

then denatured and prepared according to Illumina MiSeq guidelines using 20 pM of each the sample and PhiX control. MiSeq reagent kit 2 x 150 bp v2 was used for sequencing of E. coli isolates to generate 250 nucleotides long paired end reads with approximate total output of 4.5 Gb. With E. coli sized genomes, the average coverage if one bacterial genome was run on one cell would be 978 times (4500000000/4600000). Therefore, to obtain the minimum level of coverage of 30 time, (978/30), 32 samples can be run on with one v2 kit.

77 isolates were obtained from the Severn Trent catchment (16 CSO impacted site, 61 combined CSO and WWTP effluent impacted site). An initial 30 DNA samples were run on the Miseq at University of Technology, Sydney. The remaining 47 samples were run on the Hiseq at the Garvan institute of medical research.

2.22.1 Generating rough assemblies

Raw reads were assembled using the revised A5-miseq de novo assembly pipeline (Coil, 2015) which consists of five steps: 1) read cleaning, 2) contig assembly, 3) crude scaffolding, 4) misassembly correction, 5) final scaffolding. Assembly summary statistics and base-call estimates are produced at stage 5. Rough assemblies were carried out at UTS by Professor Aaron Darling.

Preliminary annotations were generated using the automated annotation software RAST (Overbeek et al., 2014, Aziz et al., 2012). Annotation of ARG was performed using the Resistance Gene Identifier (RGI) Version 2 on the Comprehensive Antibiotic Resistance Database (CARD) website (McArthur et al., 2013) Contigs were also run through CARD database to look for any resistance genes missed by the antibiotic resistance database.

2.22.2 Phylosift analysis

Phylogenetic analysis of genomes was performed using PhyloSift (Darling et al., 2014) which is based on phylogenetically correlating based on 37 “elite” marker gene families that have largely congruent phylogenetic histories. These genes represent ~1% of an average bacterial genome. Phylosift results were to generate phylogenetic trees using FastTree and were analysed using FigTree (found at

http://tree.bio.ed.ac.uk/people/) which is a graphical viewer for phylogenetic trees. Trees with and without references strains were performed (Table 2.7).

Table 2.7 Reference strains. Taken from the GenBank database used for comparative genomics

Description Accession number

E. coli 042 (EHEC) FN554766.1

E. coli 536 (UPEC) CP000247

E. coli O26:H11 11368 (EHEC) AP010955

E. coli O55 H7 CB9615
 CP00184 CD306 ST131 CP013831.1 CFT073 AE014075.1 E2348/69 NC_011601.1 E24277A CP000800.1 EC958 ST131 NZ_HG941718.1

EDL933 O157:H7 (EHEC) NZ_CP008957.1

JJ1886 ST131 CP006784.1 JJ1897 ST131 _CP013837.1 K12 MG1655 U00096.3 LF82 (AIEC) CU651637.1 NA114 ST131 _{NC_017644.1} O26:H11 AP010953.1

O103:H2 (EHEC) AP010958.1

O111:H- (EHEC) AP010960.1

O157:H7 Sakai BA000007.2

E. coli O103:H2 str. 12009 plasmid pO103 PO103 AP010959.1 RM12579 O55:H7 (EPEC) _CP003109.1 S. boydii ATCC_9210 CP011511.1 S. flexineri 2a str. 301 AE005674.2 S. flexneri 2a str. 2357T NC_004741.1 S. sonnei FORC_011 NZ_CP010829.1 UMNK88 (ETEC) CP002733.1 UTI89 (UPEC) NC_007946/1 W3110 K12 AP009048.1 ZH063 ST131 NZ_CP014522.1

2.22.3 Comparative genomic and MLST analysis

The move contigs function in MAUVE (Darling et al., 2011) was used for whole genome alignment to reference genomes (Table 2.7). It identifies conserved segments which appear free from genome rearrangements. These regions are referred to as Locally Collinear Blocks (LCBs), the fewer the LCBs, the better the alignment. The best alignments were chosen based on the highest weight score which is a measure of how likely the predicted rearrangement is and whether it in fact exists. Scaffolds unable to align with the reference sequence are referred to as the “accessory genome”. Scaffolds identified in MAUVE as regions of interest were further investigated using BLASTn® searches were carried out against databases made at University of Technology, Sydney for plasmid possession, virulence genes, antibiotic resistance genes, phylogrouping, O- and H-antigen typing to characterise strains. Progressive MAUVE alignments were performed between strains of the same sequence time using the arranged contig FASTS files generated using the move contigs function. PubMLST (http://pubmlst.org/) was used to identify sequence types of Escherichia isolates using the Achtman E. coli MLST scheme (Achtman et al., 2012) (http://mlst.warwick. ac.uk/mlst/). BRIG (Alikhan, 2011) was used for visual comparisons of resistance and virulence genes between key strains.

2.22.4 Investigating virulence, antimicrobial resistance and plasmid incompatibility type

BLAST searches against a databases of virulence genes, AMR genes and plasmid incompatibility groups were carried out using BLAST databases provided by UTS.

2.22.5 Toxin-antitoxin searching

TA finder was used to search the TADB (Shao et al., 2011) was used to determine whether strains possessed toxin-antitoxin systems to help elucidate whether

persistence formation was involved in the mechanisms behind significant E. coli