V.K. Petrova 1, C.W. Donnelly 1, G.A. Petrucci 2, K.A. Puzey 3 ¹Department of Animal Sciences,...
1
V.K. Petrova 1 , C.W. Donnelly 1 , G.A. Petrucci 2 , K.A. Puzey 3 ¹Department of Animal Sciences, Nutrition and Food Sciences, 2 Department of Chemistry, The University of Vermont Burlington, Vermont 05405 3 QuantaSpec Inc., Essex Junction, VT 05452 Fourier Transform Infrared (FTIR) spectroscopy provides a rapid, highly selective and reproducible means for the chemically based discrimination of intact microbial cells which make the method valuable for large-scale screening of foods. Unlike existing biological technologies that are both labor intensive and expensive due to the high volume of reagents and supplies required for strain identification, FTIR spectroscopy has a great potential in the food industry as a rapid and cost-effective foodborne pathogen identification method. INTRODUCTION Effect of growth media. Selected strains were grown in various selective media. L. innocua CWD 142 and L. welshimeri FSL H6-O17 were grown in the University of Vermont Broth (UVM); E.coli K12 HB 101 and E.coli ATCC29181 in Escherichia coli (EC) broth; S. cholerasuis ATCC 13314 and S. subterranea ATCC BAA-836 in Selenite Cystine (SC) Broth and E. sakazakii ATCC 29544 and E. aerogenes ATCC 13048 in EC broth with addition of novobiocin (0.02 g/l) (mEC broth). Each bacterial culture was grown at 37Cº for 18 h, centrifuged and washed in Butterfield's phosphate buffer. Two sampling replications for each bacterial suspension were deposited on membrane cards at levels of 10 8 – 10 9 CFU/ml, dried under laminar flow at room temperature for 3-4 h and then analyzed with Thermo Nicolet IR200 FTIR spectrometer (Madison, WI). Bacterial spectra was then compared against spectra obtained from bacteria grown in general TSB. Effect of growth phase. L. innocua CWD 142 and L. welshimeri FSL H6-017 were grown in TSB media at 37°C with agitation to reach logarithmic or stationary phase. Five ml aliquots of cultures were removed from each tube, centrifuged, washed, and placed on the membrane filters in duplicate. The filters were dried before the FTIR spectra analysis. Effect of genotypes. Listeria strains were characterized by ribotyping with the RiboPrinter Microbial Characterization System (Qualicon, Wilmington, Del) using EcoR1 chromosomal digests to confirm that they were of unique identity. L. innocua Ribotype 1-1910-7, L. innocua Ribotype 5-415-5, L. welshimeri Ribotype 1-1921-4, and L. welshimeri Ribotype 2-864-3 were used in these experiments. These cultures were grown in TSB media at 37Cº. Sample preparation and FTIR spectra analysis were performed as described earlier. Effect of food matrices. Retail commercial samples of nonfat dry skim milk powder (NDSMP) and powdered infant formula (PIF) were aseptically prepared according to the manufacturer’s instructions. Overnight cultures of E. sakazakii ATCC 29544 and E. aerogenes ATCC 13048 were inoculated into 9 ml tubes of NDSMP and PIF (10 tubes per culture) and grown at 37Cº for an 18 h incubation period. Sample preparation and FTIR analysis were performed as outlined previously. Effect of sanitizing chemicals. Commonly used food processing plant environmental sanitizers including sodium hypochlorite (XY-12, Ecolab, St. Paul, Minn), quaternary ammonium compound (Ster-Bac, Ecolab), iodine based sanitizer (Mikroklene DF, Ecolab, St. Paul, Minn.), and Ethylene Diaminetetraacetic Acid (EDTA) (Thermo Fisher Scientific, Waltham, MA) were evaluated as chemical interferents during bacterial spectral signature analysis. One ml of an 18 h culture of L. innocua CWD 142 and L. welshimeri FSL H6-017 grown in TSB at 37Cº were exposed for 2 and 3 min to 50 and 200 ppm of sodium hypochlorite and quaternary ammonium compound, respectively; 6 and 25 ppm of iodine based sanitizer; and 50 and 300 ppm of EDTA to reflect concentrations commonly used in food plant environments and food preservation. Samples were then centrifuged, washed and analyzed as described previously. MATERIALS and METHODS Data analysis was performed using multivariate discriminant analysis of the first derivative of the FTIR spectral data. ATR and transmission data were analyzed separately. Linear combinations of the spectra, or canonical variates, were formed to maximize the distance between two or more predefined group means and to provide the best discrimination of the groups. The groups were scaled so that the distances between them represent the Mahalanobis distances, Mij, which is the distance between groups in within-group standard deviation units. Canonical scores for the two most significant variates were used to create a graphical two-dimensional representation of the data. Bacterial Identification Testing. Identification algorithms (Trident™ QuantaSpec) were developed by QuantaSpec Inc. (Essex Junction, VT) DATA ANALYSIS • Representative bacterial FTIR spectra collected using transmission and ATR data are shown in Fig. 1 (A, B), respectively. The cross validation testing showed successful differentiation between studied bacteria, resulting in 100% discrimination among these cultures. One hundred fifty nine spectra, collected using the ATR approach, were correctly assigned to their species/strain group, and 157 spectra, collected with transmission approach, were correctly assigned to their species/strain group (Table 1). • The multivariate approach demonstrated the ability of infrared spectroscopy to successfully discriminate among cultures of all tested bacterial species. Graphical differentiation of the spectra is shown in Fig. 2. All tested bacterial species were separated into eight separate clusters corresponding to L. innocua CWD 142, L. welshimeri FSL H6-O17, E. coli K12 HB 101, E. coli ATCC 29181, S. cholerasuis ATCC 13314, S. subterranea ATCC BAA-836, E. sakazakii ATCC 29544, and E. aerogenes ATCC 13048. The method demonstrated very good reproducibility as evidenced by the tight clustering of specific bacterial samples grown in separated tubes on the same day and when repeated on subsequent days. The ATR data analysis resulted in Gram-negative E. coli K12, E. coli ATCC 29181, S. cholerasuis, S. subterranean, E. sakazakii, and E. aerogenes bacterial clusters which grouped closely together, and may be attributed to genetic similarities shared by members of the same genera Enterobacteriaceae (Fig. 2). • It was found that heterogeneity between spectra from L. innocua CWD 142 and L. welshimeri FSL H6-017 grown to logarithmic and stationary phases was not significant to influence the discrimination between the strains (Table 1). • FTIR approach was 100% successful in differentiation between genetically similar bacteria (Table 1). Multivariate discriminant helped clearly discriminate L. innocua and L. welshimeri ribotypes (Fig. 5). • The cross validation approach showed that the separation of bacterial species was minimally affected by variations in broth medium, sanitizer and food matrices (Table 1). FTIR spectroscopy has proven to be sensitive and highly valuable in differentiation of bacterial cells at the genus, species, and even subspecies levels. This technique may be very useful in epidemiological studies following a foodborne illness outbreak, because of its ability to provide a rapid means of assessing the biochemical properties of bacterial isolates. Because the accuracy of the method is not affected by the addition of chemicals or food matrices, it might serve as a powerful and economical tool for monitoring the safety of the food supply and potential sources of environmental contamination. Such advantages of the method as its cost- and time-effectiveness, requirement for no sample pretreatment or extensive sample standardization make it a good alternative to existing detection methods. CONCLUSIONS 1. Al-Holy MA, Lin M, Al-Qadiri H, Cavinato AG, Rasco BA. 2006. Classification of foodborne pathogens by Fourier Transform infrared spectroscopy and pattern recognition techniques. J Rapid Methods Auto Microbiol. 14:189-200. 2. Helm D, Labischinski H, Schallehn G, Naumann D. 1991. Classification and identification of bacteria by Fourier- transform infrared spectroscopy. J Gen Microbiol. 137(1):69-79. 3. Rodriguez-Saona LE, Khambaty FM, Fry FS, Calvey EM. 2001. Rapid detection and identification of bacterial strains by Fourier Transform Near-Infrared Spectroscopy. J Agric Food Chem. 49(2):574-9. 4. Udelhoven T, Naumann D, Schmitt J. 2000. Development of a hierarchical classification system with artificial neural networks and FT-IR Spectra for the identification of bacteria. Appl Spectrosc. 54:1471-9. Although application of FTIR for bacterial differentiation was reported in the past (Helm and others 1991; Udelhoven and others 2000; Rodriguez-Saona and others 2001; Al-Holy and others 2006), such efficacious results have not yet been documented. Our analysis indicated that superior bacterial classification rate can be achieved by minimization of the variability in the FTIR spectra and normalization of the spectral data by application of the novel means of analysis (Trident™ QuantaSpec) and by reduction of the number of data points to those that carry the most significant differentiation power. DISCUSSION The goals of the present study were to assess the effect of food matrices, different sanitizing compounds and growth media, on the ability of the method to accurately identify and classify L. innocua, L. welshimeri, E. coli, S. cholerasuis, S. subterranea, E. sakazakii, and E. aerogenes. Moreover, use of FTIR spectroscopy for discrimination of L. innocua and L. welshimeri of different genotypes and the effect of growth phase on identification accuracy of L. innocua and L. welshimeri were tested. Moreover, we assessed the sensitivity of FTIR to cells in various growth stages and its ability to differentiate bacteria within the same genotype. OBJECTIVES TABLE 1. Identification rate of pathogen simulates as affected by growth phase, genotype, growth media and chemical interferents. FIG 1. Representative spectra of L. innocua, L. welshimeri, E. coli, S. cholerasuis, S. subterranea, E. sakazakii, and E. aerogenes collected using Transmission FTIR (A) and ATR (B) data REFERENCES Impact of Growth Phase, Chemical and Food Matrices on Differentiation of Foodborne Pathogens by FTIR Spectroscopy -0.1 -0.05 0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 0.4 0 500 1000 1500 2000 2500 3000 3500 4000 4500 L.inocua L.welsh E.coli K12 E.coli E.aerog E.sak S.chol S.subter Li - L. innocua, Lw - L. welshimeri, EcA - E. coli ATCC 29181, EcK - E.coli K12, Sc - S. cholerasuis, Ss - S. subterranea, Es - E. sakazakii Ea - E. aerogenes -1.0 -0.5 0.0 0.5 1.0 Canonical2 Li Li-a Lw Lw-a -50000000 0 50000000 100000000 Canonical1 Li - L. innocua Ribotype 1-1910-7 Li-a - L.inocua Ribotype 5-415-5 Lw – L. welshimeri Ribotype 1-1921-4 Lw-a - L. welshimeri Ribotype 2-864-3 FIG 2. Bacterial Discrimination using multivariate discriminant analysis – ATR FTIR data. Aim of the study Strains used ATR Transmission Differentiation of Bacteria L. innocua CWD 142, L. welshimeri FSL H6-O17 E. coli K12 HB 101 E. coli ATCC 29181 S. cholerasuis ATCC 13314 S. subterranea ATCC BAA- 836 E. sakazakii ATCC 29544 E. aerogenes ATCC 13048 159/159 (100%) 157/157 (100%) Effect of Growth Phase L. innocua CWD 142, L. welshimeri FSL H6-O17 80/80 (100%) 80/80 (100%) Effect of Genotype L. innocua Ribotype 1- 1910-7 L.inocua Ribotype 5-415- 5 L. welshimeri Ribotype 1- 1921-4 L. welshimeri Ribotype 2- 864-3 80/80 (100%) 80/80 (100%) Effect of Growth Media L. innocua CWD 142, L. welshimeri FSL H6-O17 E. coli K12 HB 101 E. coli ATCC 29181 S. cholerasuis ATCC 13314 S. subterranea ATCC BAA- 836 E. sakazakii ATCC 29544 E. aerogenes ATCC 13048 316/319 (99.1%) 317/319 (99.4%) Effect of Iodophor Sanitizer L. innocua CWD 142, L. welshimeri FSL H6-O17 80/80 (100%) 80/80 (100%) Effect of Quarternay Ammonia L. innocua CWD 142, L. welshimeri FSL H6-O17 80/80 (100%) 80/80 (100%) Effect of Chlorine sanitizer L. innocua CWD 142, L. welshimeri FSL H6-O17 79/79 (100%) 80/80 (100%) -0.2 -0.1 0 0.1 0.2 0.3 0.4 0.5 0 500 1000 1500 2000 2500 3000 3500 4000 4500 L.inocua L.welshimeri E.coli K12 E.coli E. aer. TSB E.sak S.chol S. subter A B -5000 -4000 -3000 -2000 -1000 0 1000 2000 3000 Canonical2 Ea EcA EcK Es Li Lw Sc Ss -200000 0 100000 200000 300000 400000 500000 600000 700000 800000 900000 Canonical1 FIG 3. Discrimination of L. innocua and L. welshimeri ribotypes using multivariate discriminant analysis - ATR FTIR data. RESULTS
-
Upload
jane-gardner -
Category
Documents
-
view
216 -
download
1
Transcript of V.K. Petrova 1, C.W. Donnelly 1, G.A. Petrucci 2, K.A. Puzey 3 ¹Department of Animal Sciences,...
V.K. Petrova1, C.W. Donnelly1, G.A. Petrucci2, K.A. Puzey3
¹Department of Animal Sciences, Nutrition and Food Sciences, 2Department of Chemistry, The University of Vermont Burlington, Vermont 054053QuantaSpec Inc., Essex Junction, VT 05452
Fourier Transform Infrared (FTIR) spectroscopy provides a rapid, highly selective and reproducible means for the chemically based discrimination of intact microbial cells which make the method valuable for large-scale screening of foods. Unlike existing biological technologies that are both labor intensive and expensive due to the high volume of reagents and supplies required for strain identification, FTIR spectroscopy has a great potential in the food industry as a rapid and cost-effective foodborne pathogen identification method.
INTRODUCTION
Effect of growth media. Selected strains were grown in various selective media. L. innocua CWD 142 and L. welshimeri FSL H6-O17 were grown in the University of Vermont Broth (UVM); E.coli K12 HB 101 and E.coli ATCC29181 in Escherichia coli (EC) broth; S. cholerasuis ATCC 13314 and S. subterranea ATCC BAA-836 in Selenite Cystine (SC) Broth and E. sakazakii ATCC 29544 and E. aerogenes ATCC 13048 in EC broth with addition of novobiocin (0.02 g/l) (mEC broth). Each bacterial culture was grown at 37Cº for 18 h, centrifuged and washed in Butterfield's phosphate buffer. Two sampling replications for each bacterial suspension were deposited on membrane cards at levels of 108 – 109 CFU/ml, dried under laminar flow at room temperature for 3-4 h and then analyzed with Thermo Nicolet IR200 FTIR spectrometer (Madison, WI). Bacterial spectra was then compared against spectra obtained from bacteria grown in general TSB. Effect of growth phase. L. innocua CWD 142 and L. welshimeri FSL H6-017 were grown in TSB media at 37°C with agitation to reach logarithmic or stationary phase. Five ml aliquots of cultures were removed from each tube, centrifuged, washed, and placed on the membrane filters in duplicate. The filters were dried before the FTIR spectra analysis. Effect of genotypes. Listeria strains were characterized by ribotyping with the RiboPrinter Microbial Characterization System (Qualicon, Wilmington, Del) using EcoR1 chromosomal digests to confirm that they were of unique identity. L. innocua Ribotype 1-1910-7, L. innocua Ribotype 5-415-5, L. welshimeri Ribotype 1-1921-4, and L. welshimeri Ribotype 2-864-3 were used in these experiments. These cultures were grown in TSB media at 37Cº. Sample preparation and FTIR spectra analysis were performed as described earlier. Effect of food matrices. Retail commercial samples of nonfat dry skim milk powder (NDSMP) and powdered infant formula (PIF) were aseptically prepared according to the manufacturer’s instructions. Overnight cultures of E. sakazakii ATCC 29544 and E. aerogenes ATCC 13048 were inoculated into 9 ml tubes of NDSMP and PIF (10 tubes per culture) and grown at 37Cº for an 18 h incubation period. Sample preparation and FTIR analysis were performed as outlined previously. Effect of sanitizing chemicals. Commonly used food processing plant environmental sanitizers including sodium hypochlorite (XY-12, Ecolab, St. Paul, Minn), quaternary ammonium compound (Ster-Bac, Ecolab), iodine based sanitizer (Mikroklene DF, Ecolab, St. Paul, Minn.), and Ethylene Diaminetetraacetic Acid (EDTA) (Thermo Fisher Scientific, Waltham, MA) were evaluated as chemical interferents during bacterial spectral signature analysis. One ml of an 18 h culture of L. innocua CWD 142 and L. welshimeri FSL H6-017 grown in TSB at 37Cº were exposed for 2 and 3 min to 50 and 200 ppm of sodium hypochlorite and quaternary ammonium compound, respectively; 6 and 25 ppm of iodine based sanitizer; and 50 and 300 ppm of EDTA to reflect concentrations commonly used in food plant environments and food preservation. Samples were then centrifuged, washed and analyzed as described previously.
MATERIALS and METHODS
Data analysis was performed using multivariate discriminant analysis of the first derivative of the FTIR spectral data. ATR and transmission data were analyzed separately. Linear combinations of the spectra, or canonical variates, were formed to maximize the distance between two or more predefined group means and to provide the best discrimination of the groups. The groups were scaled so that the distances between them represent the Mahalanobis distances, Mij, which is the distance between groups in within-group standard deviation units. Canonical scores for the two most significant variates were used to create a graphical two-dimensional representation of the data. Bacterial Identification Testing. Identification algorithms (Trident™ QuantaSpec) were developed by QuantaSpec Inc. (Essex Junction, VT) specifically for this study. Identification algorithms were tested using a Cross Validation approach.
DATA ANALYSIS
• Representative bacterial FTIR spectra collected using transmission and ATR data are shown in Fig. 1 (A, B), respectively. The cross validation testing showed successful differentiation between studied bacteria, resulting in 100% discrimination among these cultures. One hundred fifty nine spectra, collected using the ATR approach, were correctly assigned to their species/strain group, and 157 spectra, collected with transmission approach, were correctly assigned to their species/strain group (Table 1).
• The multivariate approach demonstrated the ability of infrared spectroscopy to successfully discriminate among cultures of all tested bacterial species. Graphical differentiation of the spectra is shown in Fig. 2. All tested bacterial species were separated into eight separate clusters corresponding to L. innocua CWD 142, L. welshimeri FSL H6-O17, E. coli K12 HB 101, E. coli ATCC 29181, S. cholerasuis ATCC 13314, S. subterranea ATCC BAA-836, E. sakazakii ATCC 29544, and E. aerogenes ATCC 13048. The method demonstrated very good reproducibility as evidenced by the tight clustering of specific bacterial samples grown in separated tubes on the same day and when repeated on subsequent days. The ATR data analysis resulted in Gram-negative E. coli K12, E. coli ATCC 29181, S. cholerasuis, S. subterranean, E. sakazakii, and E. aerogenes bacterial clusters which grouped closely together, and may be attributed to genetic similarities shared by members of the same genera Enterobacteriaceae (Fig. 2).
• It was found that heterogeneity between spectra from L. innocua CWD 142 and L. welshimeri FSL H6-017 grown to logarithmic and stationary phases was not significant to influence the discrimination between the strains (Table 1).
• FTIR approach was 100% successful in differentiation between genetically similar bacteria (Table 1). Multivariate discriminant helped clearly discriminate L. innocua and L. welshimeri ribotypes (Fig. 5).
• The cross validation approach showed that the separation of bacterial species was minimally affected by variations in broth medium, sanitizer and food matrices (Table 1).
FTIR spectroscopy has proven to be sensitive and highly valuable in differentiation of bacterial cells at the genus, species, and even subspecies levels. This technique may be very useful in epidemiological studies following a foodborne illness outbreak, because of its ability to provide a rapid means of assessing the biochemical properties of bacterial isolates. Because the accuracy of the method is not affected by the addition of chemicals or food matrices, it might serve as a powerful and economical tool for monitoring the safety of the food supply and potential sources of environmental contamination. Such advantages of the method as its cost- and time-effectiveness, requirement for no sample pretreatment or extensive sample standardization make it a good alternative to existing detection methods.
CONCLUSIONS
1. Al-Holy MA, Lin M, Al-Qadiri H, Cavinato AG, Rasco BA. 2006. Classification of foodborne pathogens by Fourier Transform infrared spectroscopy and pattern recognition techniques. J Rapid Methods Auto Microbiol. 14:189-200.
2. Helm D, Labischinski H, Schallehn G, Naumann D. 1991. Classification and identification of bacteria by Fourier-transform infrared spectroscopy. J Gen Microbiol. 137(1):69-79.
3. Rodriguez-Saona LE, Khambaty FM, Fry FS, Calvey EM. 2001. Rapid detection and identification of bacterial strains by Fourier Transform Near-Infrared Spectroscopy. J Agric Food Chem. 49(2):574-9.
4. Udelhoven T, Naumann D, Schmitt J. 2000. Development of a hierarchical classification system with artificial neural networks and FT-IR Spectra for the identification of bacteria. Appl Spectrosc. 54:1471-9.
Although application of FTIR for bacterial differentiation was reported in the past (Helm and others 1991; Udelhoven and others 2000; Rodriguez-Saona and others 2001; Al-Holy and others 2006), such efficacious results have not yet been documented. Our analysis indicated that superior bacterial classification rate can be achieved by minimization of the variability in the FTIR spectra and normalization of the spectral data by application of the novel means of analysis (Trident™ QuantaSpec) and by reduction of the number of data points to those that carry the most significant differentiation power.
DISCUSSION
The goals of the present study were to assess the effect of food matrices, different sanitizing compounds and growth media, on the ability of the method to accurately identify and classify L. innocua, L. welshimeri, E. coli, S. cholerasuis, S. subterranea, E. sakazakii, and E. aerogenes. Moreover, use of FTIR spectroscopy for discrimination of L. innocua and L. welshimeri of different genotypes and the effect of growth phase on identification accuracy of L. innocua and L. welshimeri were tested. Moreover, we assessed the sensitivity of FTIR to cells in various growth stages and its ability to differentiate bacteria within the same genotype.
OBJECTIVES
TABLE 1. Identification rate of pathogen simulates as affected by growth phase, genotype, growth media and chemical interferents.
FIG 1. Representative spectra of L. innocua, L. welshimeri, E. coli, S. cholerasuis, S. subterranea, E. sakazakii, and E. aerogenes collected using Transmission FTIR (A) and ATR (B) data
REFERENCES
Impact of Growth Phase, Chemical and Food Matrices on Differentiation of Foodborne Pathogens by FTIR Spectroscopy
-0.1
-0.05
0
0.05
0.1
0.15
0.2
0.25
0.3
0.35
0.4
0 500 1000 1500 2000 2500 3000 3500 4000 4500
L.inocua
L.welsh
E.coli K12
E.coli
E.aerog
E.sak
S.chol
S.subter
Li - L. innocua, Lw - L. welshimeri, EcA - E. coli ATCC 29181, EcK - E.coli K12, Sc - S. cholerasuis, Ss - S. subterranea, Es - E. sakazakiiEa - E. aerogenes
-1.0
-0.5
0.0
0.5
1.0
Canonic
al2
Li
Li-a
Lw
Lw
-a
-50000000 0 50000000 100000000
Canonical1
Li - L. innocua Ribotype 1-1910-7Li-a - L.inocua Ribotype 5-415-5Lw – L. welshimeri Ribotype 1-1921-4Lw-a - L. welshimeri Ribotype 2-864-3
FIG 2. Bacterial Discrimination using multivariate discriminant analysis – ATR FTIR data.
Aim of the study Strains used ATR Transmission
Differentiation of Bacteria
L. innocua CWD 142,L. welshimeri FSL H6-O17
E. coli K12 HB 101E. coli ATCC 29181
S. cholerasuis ATCC 13314S. subterranea ATCC BAA-836
E. sakazakii ATCC 29544E. aerogenes ATCC 13048
159/159 (100%) 157/157 (100%)
Effect of Growth PhaseL. innocua CWD 142,
L. welshimeri FSL H6-O1780/80 (100%) 80/80 (100%)
Effect of Genotype
L. innocua Ribotype 1-1910-7L.inocua Ribotype 5-415-5
L. welshimeri Ribotype 1-1921-4L. welshimeri Ribotype 2-864-3
80/80 (100%) 80/80 (100%)
Effect of Growth Media
L. innocua CWD 142,L. welshimeri FSL H6-O17
E. coli K12 HB 101E. coli ATCC 29181
S. cholerasuis ATCC 13314S. subterranea ATCC BAA-836
E. sakazakii ATCC 29544E. aerogenes ATCC 13048
316/319 (99.1%) 317/319 (99.4%)
Effect of Iodophor Sanitizer
L. innocua CWD 142,L. welshimeri FSL H6-O17
80/80 (100%) 80/80 (100%)
Effect of Quarternay Ammonia
L. innocua CWD 142,L. welshimeri FSL H6-O17
80/80 (100%) 80/80 (100%)
Effect of Chlorine sanitizer
L. innocua CWD 142,L. welshimeri FSL H6-O17
79/79 (100%) 80/80 (100%)
Effect of EDTAL. innocua CWD 142,
L. welshimeri FSL H6-O1780/80 (100%) 79/79 (100%)
Effect of Food MatricesE. sakazakii ATCC 29544E. aerogenes ATCC 13048
79/79 (100%) 80/80 (100%)
-0.2
-0.1
0
0.1
0.2
0.3
0.4
0.5
0 500 1000 1500 2000 2500 3000 3500 4000 4500
L.inocua
L.welshimeri
E.coli K12
E.coli
E. aer. TSB
E.sak
S.chol
S. subter
A B
-5000
-4000
-3000
-2000
-1000
0
1000
2000
3000
Canonic
al2
Ea
EcA
EcK
Es
Li
Lw
Sc
Ss
-200000 0 100000 200000 300000 400000 500000 600000 700000 800000 900000
Canonical1
FIG 3. Discrimination of L. innocua and L. welshimeri ribotypes using multivariate discriminant analysis - ATR FTIR data.
RESULTS
