| Objective To investigate the role of matrix-assisted laser desorption ionization time of flight mass spectrometry(MALDI-TOF MS)in the rapid identification of ultra-broad spectrum β-lactamase(ESBLs)produced by E.Coli and metalloenzyme and serine protease produced by Klebsiella pneumoniae.Methods Positive control group(ESBL-producing escherichia coli,carbapenemaseproducing Klebsiella pneumoniae)and negative control group(sensitive escherichia coli,sensitive Klebsiella pneumoniae)were screened from escherichia coli and Klebsiella pneumoniae which appeared in Tangshan Workers’ Hospital in the past two years.The ESBLs enzyme produced in Escherichia coli was validated by broth dilution method(CLSI recommended).Carbapenems produced by Klebsiella pneumoniae(including metalases and serinases)were confirmatory and classified by m CIM combined with e CIM assay(CLSI recommended).The experimental data were divided into three groups:ESBL-producing escherichia coli and ESBL-sensitive escherichia coli(ESBLs group);Klebsiella pneumoniae that produces metalloenzymes and klebsiella pneumoniae sensitive that does not produce metalloenzymes(hereinafter referred to as the metalloenzyme group);Klebsiella pneumoniae producing serine protease and Rabinococcus pneumoniae susceptible to non-serine protease(hereinafter referred to as serine protease group)were rapidly detected by MALDI-TOF MS.Clin Pro Tools was used to remove baseline and exclude invalid spectrum of the experimental data of the three groups.The three models of genetic algorithm(GA),supervised neural network algorithm(SNN)and fast classification method(QC)were used for statistical analysis of all the peak graphs,so as to obtain the sensitivity,specificity and characteristic peak results of each algorithm,and find out whether there are signature characteristic peaks in each group.Results The three model algorithms were different,but the results were very similar.The sensitivity of each algorithm in ESBLs group was greater than 90.0%,among which the genetic algorithm(GA)had the highest sensitivity and specificity,reaching 99.01%,and its specificity was 91.41%;The sensitivity of SNN algorithm was 96.12%,but its specificity was 88.53%.The sensitivity of QC algorithm was 94.52%,and its specificity was 89.13%.The sensitivity of each algorithm was greater than 88.0%,among which the genetic algorithm(GA)was the most sensitive,reaching 95.43%,and its specificity was90.87%;The sensitivity of SNN algorithm was 89.64%,but its specificity was 94.19%.The sensitivity of QC algorithm was 88.97%,and its specificity was 89.99%.The sensitivity of each serine proteome algorithm was greater than 91.0%,and the SNN algorithm with the highest sensitivity and specificity was 96.35%,its specificity was92.07%;The sensitivity and specificity of genetic algorithm(GA)were 93.43% and91.05% respectively.The sensitivity and specificity of QC algorithm were 91.25% and91.66% respectively.A total of 10 differential protein peaks were detected in the three groups(P<0.05),among which peaks 98,56 and 88(mass charge scores were 9128.23,5770.48 and 8371.39Da)were the most significant(P<0.05)in ESBLs group.Peaks 77,26 and 132(mass charge ratio 7125.83,4134.2 and 10761.8Da)were the significant differential protein peaks in the metalase group(P<0.05).Peaks 1,27,52 and 102(mass charge ratio being 2067.51,4156.25,5383.37 and 8311.6Da)were the significant difference protein peaks in serine protease group(P<0.05).The detection time of ESBLs enzyme and carbapenase by traditional drug sensitivity method is more than 24 hours,while the detection time of MALDI-TOF MS for each sample is only more than ten seconds.Conclusion Mal DI-TOF MS can be used to screen the difference protein spectrum peaks between ESBL-producing and non-ESBL-producing escherichia coli.It can screen the difference protein spectrum peaks between the metalase-producing and non-enzymeproducing klebsiella pneumoniae sensitive strains.It can produce filter type of serine protease and enzyme production of sensitive bacteria pneumonia klebsiella differences protein spectrum peaks,according to the characteristic protein spectrum peak of each strain,the multidrug-resistant enterobacteriaceae bacteria can be screened directly,quickly and accurately,and the preliminary anti-infection treatment plan can be provided for clinic as soon as possible.Figure 11;Table 6;Reference 145... |