Comprehensive Assessment of Subtyping Methods for Improved Surveillance of Foodborne Salmonella

High-resolution and efficient typing for the bacterial pathogen is essential for tracking the sources, detecting or diagnosing variants, and conducting a risk assessment. However, a systematic in-field investigation of Salmonella along the food chain has not been documented. This study assessed 12 typing methods, such as antimicrobial-resistance (AMR) gene profile typing, Core Genome Multilocus Sequence Typing (cgMLST), and CRISPR multi-virulence locus sequence typing (CRISPR-MVLST), to evaluate their effectiveness for use in routine monitoring of foodborne Salmonella transmission along the poultry production chain. During 2015-16, a total of 1,064 samples were collected from poultry production chain, starting from breeding farms and slaughterhouses to the markets of Zhejiang province in China. A total of 61 consecutive unique Salmonella isolates recovered from these samples were selected for genome sequencing and further comparative typing analysis. Traditional typing methods, including serotyping, AMR phenotype-based typing, as well as modern genotyping approaches, were evaluated and compared by their discrimination index (DI). The results showed that the serotyping method identified nine serovars. The gold standard cgMLST method indicated only 18 different types (DI = 0.8541), while the CRISPR-MVLST method detected 30 types (DI = 0.9628), with a higher DI than all examined medium-resolution WGS-based genotyping methods. We demonstrate that the CRISPR-MVLST might be used as a tool with high discriminatory power, comparable ease of use, ability of tracking the source of Salmonella strains along the food chain and indication of genetic features especially virulence genes. The available methods with different purposes and laboratory expertise were also illustrated to assist in rational implementation. IMPORTANCE In public health field, high-resolution and efficient typing of the bacterial pathogen is essential, considering source-tracking and risk assessment are fundamental issues. Currently, there are no recommendations for applying molecular characterization methods for Salmonella along the food chain, and a systematic in-field investigation comparing subtyping methods in the context of routine surveillance was partially addressed. Using 1,064 samples along a poultry production chain with a considerable level of Salmonella contamination, we collected representative isolates for genome sequencing and comparative analysis by using 12 typing techniques, particularly with whole-genome sequence (WGS) based methods and a recently invented CRISPR multi-virulence locus sequence typing (CRISPR-MVLST) method. CRISPR-MVLST is identified as a tool with higher discriminatory power compared with medium-resolution WGS-based typing methods, comparable ease of use and proven ability of tracking Salmonella isolates. Besides, we also offer recommendations for rational choice of subtyping methods to assist in better implementation schemes.