Towards DNA Barcoding for Rapid Pathogen Detection

Lance Palmer
Siemens Research

Next-generation sequencing technologies will soon enable researchers to sequence hundreds to thousands of bacterial strains at low cost. By exploiting this explosion of information it will be possible to develop DNA barcodes that will allow physicians to accurately detect and distinguish among bacterial strains present in a sample in just a few hours. To achieve this goal, there are hurdles which must be overcome today. Due to the small sizes of sequencing reads that are produced by these technologies, genome assembly can be problematic. We are currently working on several methods to improve sequence assembly from short reads, including methods for error detection in short reads, assembly of short reads into long read using de Bruijn graphs, and overlap correction using machine learning techniques. Once sequencing reads have been assembled we design a DNA barcode: sets of PCR primers and probes that can distinguish one set of bacteria from another. Our current work is focused on distinguishing among methicillin-resistant Staphylococcus aureus (MRSA), methicillin-sensititve S. aureus (MSSA) and coagulase negative Staphylococcus (CNS).