Clinical sample collection and bacterial strains
In this study, A total number of 91 Shigella strains including S. dysenteriae (n=18), S. flexneri (n=25), S. boydii (n=23) and S. sonnei (n=25), were used. Shigella strains were isolated from stool samples collected totally from 1862 patients, female and male aged from 5 to 75 years, with acute diarrhea who admitted to the Milad hospital, Tehran, Iran; between from September 2016 to December 2020. In this study, four species of Shigella used as the reference bacterial strains and positive controls including Sh. dysenteriae ATCC 13313, Sh. flexneri PTCC 1865, Sh. boydii ATCC 12030 and Sh. sonnei PTCC 1777 were purchased from Pasteur Institute (Pasteur In., Tehran, Iran) in lyophilized form. All lyophilized reference bacterial strains were activated by inoculation in Trypticase Soy Broth (TSB, Merck, Germany) and incubation for 18 h at 37 °C. All strains were subjected to DNA extraction before the FTIR analysis.
Isolation and identification of Shigella species from stool samples
Shigella species were isolated from stool samples and differentiated according to the conventional methods as gold standards previously described by Mokhtari et al.33 and Phiri et al.34 Sterilized disposable inoculation loop of stool samples were directly inoculated on xylose lysine deoxycholate agar (XLD, Merck, Germany) and incubated at 37 °C for 24 h aerobically. Suspected colonies including red ones on XLD agar morphologically resembling Shigella were isolated and subjected to biochemical tests. Lysine iron decarboxylase (LIA, Merck, Germany), triple sugar iron (TSI, Merck, Germany), IMViC (Oxoid Ltd., UK), and urease production (Merck, Germany) tests were used to confirm the suspected colonies and identify Shigella in stool samples. Genus and species of each presumptive Shigella isolate were serologically determined and identified by slide agglutination assay using the commercial Shigella genus and species antisera kits (Difco Co., MI, USA), respectively. All Shigella isolates were stocked in TSB (Merck, Germany) containing 20% glycerol and kept at -70 °C until further experiments.
Presumptive Shigella isolates on XLD agar from stool samples and the cultured reference Shigella strains were subjected to DNA extraction. The genomes of the isolates and strains were extracted using the Sinaclon bacterial DNA extraction commercial kit (Sinaclon Co., Tehran, Iran) according to the manufacturers’ instructions. The quantity and quality of the extracted DNA were determined using the NanoDrop 2000 spectrophotometer (ThermoFisher, USA). Also, the final concentrations of all extracted DNA templates were adjusted to 20 ng. µL-1 and they were kept at -20 °C until further analysis.
Due to the water absorption bands affect significantly the interpretation of FTIR spectra of DNA, the templates (50 µL) were dried and used for FTIR spectroscopy. KBr-FTIR technique previously described by Han et al.28 using an FTIR spectrometer (Magna 550, Madison, USA) over the wavenumber range from 4000 to 400 cm-1 was used. The spectra were generated using 64 scans with a resolution of 4 cm-1. Analysis of each DNA template was carried out in triplicates.
The FTIR spectral data were analyzed using dimension reduction and multivariate statistical methods by the SPSS version 26.0 software package for Windows (SPSS Inc., Illinois, USA) and the Minitab version 19 software (Minitab Inc., USA). In this study, principal component analysis (PCA) and hierarchical clustering analysis (HCA) methods were used to analyze the FTIR spectral data obtained from the DNA templates. The Bartlett`s test of sphericity and Kaiser-Meyer-Olkin test were performed prior to the PCA35. Twenty-five types of the analysis results were selected and analyzed for evaluation of the significant distribution with the infrared intensities and wavenumbers of the Shigella isolates spectral data. Differentiation rate of the developed method in this study was evaluated according to specificity, sensitivity, differentiation error and correct differentiation rate of the assay, which were calculated as follows:
Specificity = TN/(TN+FP) × 100
Sensitivity = TP/(TP+FN) × 100
Differentiation error = (FP+FN)/(TP+TN+FP+FN) × 100
Correct differentiation rate = 1 - (FP+FN)/(TP+TN+FP+FN) × 100
where the true positive (TP) indicates the number of correctly identified species presented in samples, the false positive (FP) indicates the number of false identified species presence in samples, the true negative (TN) indicates the number of correctly identified species not presented in samples and the false negative (FN) indicates the number of false identified species not presented in samples28. All measurements were carried out in triplicates.