Robert-von-Ostertag-Str. 7-13
14163 Berlin
+49 30 838 51843 / 66949
mikrobiologie@vetmed.fu-berlin.de
Aim:
We investigated different sequencing and assembling methods regarding their ability to detect resistance genes and predict the association to mobile genetic elements as plasmids in E. coli.
Methods:
Five E. coli isolates with distinct resistance profiles and diverse plasmid contents were characterized in detail by antimicrobial susceptibility testing, XbaI-/ S1-PFGE profiling, and filter mating studies. The biological characteristics of the isolates were compared with in silico predictions using different sequencing strategies. Raw reads were generated using the Illumina NextSeq, the long-read systems PacBio Sequel and MinION. The reads were used individually or in combinations to assess appropriate workflows for reliable in silico-based typing purposes.
Results:
Long-reads alone were shown to be error-prone, affecting the annotation of genes and the prediction of small plasmids. However, their use represented the best option to generate genome-scaffolds. Short-reads were reliable for detecting resistance genes, but were not suitable for linking specific resistance determinants to the chromosome or to plasmid types. Moreover, short-read sequencing missed duplicated resistance genes. The use of hybrid-assemblies led to the best consensus between in silico and in vitro results, as all duplications and small plasmids were predicted and provided closed plasmid for allmost all extrachromosomal elements.
Conclusions:
Hybrid assembling provided a thorough overview on extrachromosomal DNA elements, associated with antimicrobial resistances and detailed insights into the genomes genetic composition. Overall, it will be worth extending the routine sequence diagnostic to hybrid sequencing, when a reference-grade complete bacterial genome is aimed, or extrachromosomal structures needs to be fully understood.