Heather Carleton from the Centers for Disease Control & Prevention (CDC) presented at London Calling 2024. The session was titled “Whole-genome sequencing in PulseNet foodborne molecular surveillance systems.” They began discussing the impact of foodborne illness in the United States and globally. Carleton noted that in some cases, hospitalization and death occur. This impact affects morbidity, mortality, and the economy. Carleton described the foodborne surveillance system in the United States. Fecal samples are used for pathogen identification and molecular typing. Clinical labs with sequencing capacity then transfer the information to PulseNet. Close molecular matches within a timeframe are addressed by epidemiologists to search for the source of potential outbreaks. The PulseNet system has generated over half a million isolate sequences. The PulseNet system also works with PulseNet international to enhance lab capacity and surveillance. Data generated was mostly short read. With the launch of the Q20+ chemistry in 2023, the PulseNet team wanted to revisit the use of long-reads for core genome multilocus sequence typing (cgMLST). The team then compares core genome signatures. Carleton explained that they still see allele differences vs. current-validated PulseNet protocols. Why were they seeing so many differences? With updates to Dorado, they were able to reduce the number of differences and align better with the currently validated system. The team then asked how does bacterial methylation base calling align with the currently validated system with Illumina sequencing? Carleton shared heatmaps that compared isolates to the current “gold standard” and different outbreak sources. Carleton discussed examples of E. coli 0111:NM Shiga toxin-producing isolate detection. One intriguing case was a Salmonella serovar Oranienburg that with the sequencing technology again placed this isolate outside the outbreak cluster. Carleton ended by noting that they are working on new protocols, training, and updates for AMR detection, characterization, and strain-level typing. I find this so interesting and wonder how typing will change with methylation detection.
