Research On Construction And Accuracy Of Enterococci Fingerprint Database Of Antibiotic Resistance

As a cost-effective for fecal contamination sources tracking tool, Antibiotic resistance analysis (ARA) has been applied in Europe and the United States successfully. The principle is that using the same tracer with host specificity of antibiotic resistance from the feces of different source to identify fecal contamination sources. Based on the feasibility study of Enterococcus as tracing factor of ARA, it is detected 500 strains of Enterococci antibiotic resistance from five species as people, dogs, chickens, cows, and pigs. Then we used discriminant analysis to construct fingerprint database of antibiotic resistance, and obtained fingerprint database accuracy. Finally it is discussed the factors of the accuracy and promoted the application of ARA in faecel source tracking in our country.The main results as followed:(1)Enterococci of different sources fecal and strains of Enterococci standard in simulated natural water environment survival decreases with time gradually, survival time was from 18 to 20 days, but there was little difference in the survival of law; Enterococci antibiotic resistance had no change and consisted with the original condition. Therefore Enterococcus is chosen as the ARA tracing factor.(2)Analyzing the resistance rate and the multi-resistance characteristics of 500 strains of Enterococci, it is showed that antibiotic resistance had significant differences among the five species, and the antibiotic resistance can be used for constructing the Fingerprint Database. According to the resistance phenotypes of five species source with overlapping statistical, it is found that human have more chance to get overlapped phenotypic than the other species,and this may cause lower accuracy of human sources in fingerprint database.(3)Fingerprint database of antibiotic resistance was stable in any kind of discrimination mode with discriminant analysis. Under discrimination mode of Jinzhou Bay vs Xinghai Bay and human vs nonhuman, we got highest accuracy, respectively the ARCC was 71.0% and 69.0%. Validated-sets came to determine the optimal mode of the database was source vs non-human sources of man.(4)The accuracy of fingerprint database remains the same when the random sampling is large than 80%, with ARCC between 57.8% and 58.2%, sample size will not affect this fingerprint database. In the future, we will consider to use more the types of antibiotics, and remove Vancomycin, Chloramphenicol and Fosfomycin, it will increase the accuracy of the fingerprint database.