[Evaluation of effect based on different typing methods in Escherichia coli]

Objective: To evaluate the typing and clinical application effect based on clustered regularly interspaced short palindromic repeats (CRISPRs), serotype, and Multilocus Sequence Typing (MLST). Methods: The spacers, serotype and sequence type (ST) were obtained with CRISPRsFinder, SeroTypeFinder and MLST. PCR was used to amplify the CRISPRs, and the spacers were used to predict serotype and ST, then comparing with the serotype and ST. Results: We defined the I-E CRISPR/Cas as CT-Ⅰ, I-F CRISPR/Cas as CT-Ⅱ, and only CRISPR3-4 as CT-Ⅲ. We designated each unique arrangement spacer profile as a unique CRISPRs type. A total of 79 CT types, 76 serotypes, and 66 STs were identified. The CRISPRs typing was the most discriminating, with the Simpson index of 0.936, having the highest correlation with serology with the adjusted Rand index of 0.908. The CRISPRs type could divide the same serotype (ST) into two subtypes [O157∶H7(ST11), O104∶H4(ST678), and O26∶H11(ST21)]. The detection rates of CRISPR1, CRISPR2, CRISPR3, CRISPR4, and CRISPR3-4 were 81.1%, 94.5%, 1.4%, 1.4%, and 4.6%, with the accuracy rate of 95.0% and 100.0% according to the spacers to forecast O157∶H7 (ST11) and ST131. Conclusion: Based on the CRISPRs spacer, this method can be used as an essential molecular typing for E.coli, as it presents a good typing and clinical application effect.