Biophysical Society Thematic Meeting | Stockholm 2022

Physical and Quantitative Approaches to Overcome Antibiotic Resistance

Tuesday Speaker Abstracts

HIGH-RESOLUTION BACTERIAL TYPING USING OPTICAL DNA MAPPING FOR DIAGNOSING BACTERIAL INFECTIONS My Nyblom 1 ; Anna Johnning 2 ; Karolin Frykholm 1 ; Zahra Abbaspour 1 ; Marie Wrande 3 ; Albertas Dvirnas 4 ; Tobias Ambjörnsson 4 ; Christian Giske 5 ; Linus Sandegren 3 ; Erik Kristiansson 2 ; Fredrik Westerlund 1 ; 1 Chalmers University of Technology, Biology and Biological Engineering, Gothenburg, Sweden 2 Chalmers University of Technology, Mathematical Sciences, Gothenburg, Sweden 3 Uppsala University, Medical Biochemistry and Microbiology, Uppsala, Sweden High-resolution identification of bacteria is important for diagnosing bacterial infections, but can be challenging with existing methods. We use Optical DNA Mapping (ODM) to identify bacteria in clinical isolates with sub-species resolution. We have previously demonstrated that ODM can be used to accurately identify bacteria in a sample on the species level [1]. We here demonstrate that the method is applicable for high-resolution typing on the sub-species level for E. coli, K. pneumoniae and S. pyogenes. Pathogenic bacteria from clinical samples are enclosed in agarose plugs to keep DNA molecules intact during the cell lysis. Large (>100 kb) DNA molecules are extracted and a single step competitive binding-based labelling is applied to create a sequence specific emission intensity profile along the DNA. To record the intensity profile, the DNA is stretched in nanofluidic channels and imaged using fluorescence microscopy. Experimental intensity profiles are compared to a reference database, and bacteria present in the sample are identified based on discriminatively matching profiles. Bioinformatics tools are used to create a phylogenetic tree, based on which typing with taxonomic resolution higher than species can be performed. For E. coli we have a true positive rate close to 100 % at a taxonomic resolution where the E. coli species is divided into 89 different groups. This means that we can efficiently identify particularly pathogenic E. coli, such as the endemic ST131. Similar performance was observed also for K. pneumoniae and S. pyogenes. The method can also efficiently identify and characterize bacteria. By including a Cas9 restriction step, the plasmid carrying a specific (resistance) gene can be identified. Since the method is applicable directly to clinical samples, such as urine, and is very efficient in identifying bacteria in complex mixtures, we foresee that it can be an important future diagnostic tool. 4 Lund University, Astronomy and Theoretical Physics, Lund, Sweden 5 Karolinska Institute, Laboratory Medicine, Stockholm, Sweden

18