Development of a Bioinformatics Pipeline to Identify Antibiotic Resistance Biomarkers of Bacterial Pathogens

Neonatal sepsis is a severe infection with a high mortality rate of newborns worldwide. Currently, standard treatment of bacterial sepsis includes antibiotic prescription. However, the rapid evolution of antibiotic resistance in sepsis pathogens foreshadows the urgent need for new therapies. Therefore, this research focused on the development of a software pipeline allowing in silico analysis of bacterial pathogens, to identify antibiotic resistance biomarkers. Biomarkers are useful for rapid establishment of antibiotic resistance genes and mechanisms, allowing prescription of an appropriate antibiotic that would greatly reduce neonatal sepsis mortality rates.
To develop an antibiotic resistance biomarker pipeline, sequenced genomes of two diverse bacterial agents of sepsis, Gram-positive L. monocytogenes EGD-e and Gram-negative S. enterica Typhi 14028, were studied in silico. The input bacterial genome data were sequenced with an Illumina MiSeq desktop sequencer and stored in FASTQ file format. To analyse the bacterial genomes, several bioinformatics tools were reviewed and command-line tools were preferred due to their high potential for automated processes.
The overall biomarker identification pipeline developed here, was broken into two consecutive but separate parts comprising ‘Genome assembly and annotation’ and ‘Comparative genomics and biomarker identification’ sub-pipelines. The function of the first sub-pipeline included the assembly and annotation of a bacterial genome from raw sequence data. The second sub-pipeline provided a homology-based comparative genomics analysis, which required several annotated genomes as input. The comparative analyses incorporated several approaches such as phylogenetic tree construction, resistome profile investigation and pan-genome study. Biomarkers were discovered through the above techniques, after comparison of the pathogens with each other as well as with non- pathogenic strains. The intelligence on the genetic secrets of a pathogen could be utilised by a clinician or a researcher to improve the existing treatments or to develop more effective therapeutic strategies. The biomarker identification pipeline possesses the potential for the study of any bacterial pathogen. However, further development is required for full automation of this pipeline.