NF VALIDATION Validation of alternative analytical methods Application in food microbiology. Summary report

ACCREDITATION N 1-0144 PORTEE DISPONIBLE SUR WWW.COFRAC.FR Eurofins GeneScan GmbH Engesserstraße 4 D-79108 Freiburg im Breisgau GERMANY NF VALIDATION Validation of alternative analytical methods Application in food microbiology Summary report EN ISO 16140 validation study of the method for Salmonella spp. detection Qualitative method This report includes 134 pages, with 13 appendixes. Only copies including the totality of this report are authorised. Competences of the laboratory are certified by COFRAC accreditation for the analyses marked with symbol. Version 0 July 15, 2015 ADRIA DEVELOPPEMENT Creac h Gwen - F. 29196 QUIMPER Cedex - Tél. (33) 02.98.10.18.18 - Fax (33) 02.98.10.18.08 E-mail : adria.developpement@adria.tm.fr - Site web : http://www.adria.tm.fr ASSOCIATION LOI DE 1901 - N SIRET 306 964 271 00036 - N EXISTENCE 532900006329 - N TVA FR4530696427100036

1 INTRODUCTION 5 2 METHODS PROTOCOLS 5 2.1 Alternative method 5 2.2 Reference method 6 3 METHOD COMPARISON STUDY 6 3.1 Relative accuracy, relative specificity and relative sensitivity 6 3.1.1 Number and nature of samples 6 3.1.2 Artificial contamination of samples 8 3.1.3 Confirmation protocols 10 3.1.4 Test results 10 3.1.5 Calculation of relative accuracy (AC), relative sensitivity (SE) and relative specificity (SP) 15 3.1.6 Analysis of discordants 16 3.1.7 Confirmations 18 3.1.8 BPW storage for 72 h at 2 8 C 19 3.2 Relative detection level 20 3.2.1 Matrices 20 3.2.2 Contamination protocol 20 3.2.3 Results 21 3.2.4 Conclusion 21 3.3 Inclusivity / exclusivity 22 3.3.1 Test protocols 22 3.3.2 Results 22 3.3.3 Conclusion 22 3.4 Practicability 23 ADRIA Développement 2/134 July 15, 2015

4 INTER-LABORATORY STUDY ORGANISATION AND RESULTS 25 4.1 Study organisation 25 4.2 Experimental parameters control 26 4.2.1 Contamination level before inoculation, levels obtained after the artificial contaminations of the samples 26 4.2.2 Logistic conditions 27 4.2.3 Conclusion 28 4.3 Results analysis 28 4.3.1 Aerobic mesophilic flora enumeration 28 4.3.2 Expert lab results 28 4.3.3 Collaborator lab results 28 4.4 Results interpretation 30 4.4.1 Specificity and sensitivity for each method 30 4.4.2 Relative accuracy (AC) 31 4.4.3 Discordant results 32 4.5 Interpretation 32 4.5.1 Comparison of the relative accuracy, specificity and sensitivity values 32 4.5.2 Accordance (DA) 32 4.5.3 Concordance 33 4.5.4 Odds Ratio (COR) 33 5 CONCLUSION 34 Appendix 1 Flow diagrams of the alternative method 35 Appendix 2 ISO 6579:2002: Microbiology of food and animal feeding stuffs - Horizontal method for the detection of Salmonella spp. 38 Appendix 3 Artificial contamination of the samples 40 Appendix 4 - Relative accuracy, relative specificity and relative sensitivity: raw data 52 Appendix 5 - Calculation of relative accuracy (AC), relative sensitivity (SE) and relative specificity (SP) 84 Appendix 6 - Relative detection level: raw data 86 Appendix 7 - Inclusivity / exclusivity: raw data 98 Appendix 8 Results obtained by the Expert Laboratory 109 Appendix 9 Results obtained by each Collaborator 110 Appendix 10 Specificity and sensitivity 127 Appendix 11 Paired results of the alternative and reference methods for each level 128 Appendix 12 Accordance 129 Appendix 13 Concordance 131 ADRIA Développement 3/134 July 15, 2015

Before comment Quality Assurance documents related to this study can be consulted upon request from Eurofins GeneScan. The technical protocol and the result interpretation were realised according to the EN ISO 16140 and the AFNOR technical rules. Company: Eurofins GeneScan GmbH Engesserstraße 4 D-79108 Freiburg im Breisgau (Germany) Expert Laboratory: ADRIA Développement ZA Creac h Gwen F-29196 QUIMPER Cedex (France) Studied method: method for Salmonella spp. detection Validation standard : ISO 16140 (October 2003) : Food microbiology Protocol for the validation of alternative methods Reference method : ISO 6579 (2002): Horizontal method for the detection of Salmonella spp. Scope : Food products Feed products including pet food (25 g test portion) Environmental samples Pet food (375 g test portion) Milk powders, infant formula with and without probiotics (375 g test portion) Certification body: AFNOR Certification Analyses performed according to the COFRAC accreditation ADRIA Développement 4/134 July 15, 2015

1 INTRODUCTION The method for Salmonella spp. detection was validated in March 2015 in food products (Certificate number EGS 38/01 03/15). Two extension studies were performed to extend the scope: - Feed products including pet foods (25 g test portion), environmental samples, and pet foods (50 g - 375 g test portion) in May 2015; - Milk powders, infant formula with and without probiotics (50 g - 375 g test portion) in July 2015. 2 METHODS PROTOCOLS 2.1 Alternative method Principle The Salmonella detection is based on a molecular method. Protocols (See Appendix 1) The protocols are the following: Food products (except dairy products) Dairy products Environmental samples Pet foods (375 g test portion) Feed products including pet food (25 g test portion) Milk powders, infant formula with and without probiotics (375 g test portion) Enrichment step in BPW for 20 h ± 4 h at 37 C ± 1 C Enrichment step in BPW for 21 h ± 3 h at 41.5 C ± 1 C 25 g + 225 ml BPW 1 swab + 10 ml BPW 1 wipe or sponge + 225 ml BPW Incubation 21 h ± 3 h at 37 C ± 1 C 375 g + 3 375 ml pre-warmed BPW Incubation 21 h ± 3 h at 37 C ± 1 C 25 g + 225 ml BPW Incubation 21 h ± 3 h at 37 C ± 1 C 375 g + 3 375 ml pre-warmed BPW (double strength BPW for milk powders with probiotics) Incubation 21 h ± 3 h at 37 C ± 1 C ADRIA Développement 5/134 July 15, 2015

- DNA extraction on 10 µl of enrichment broth - PCR on 5 µl DNA extract on 2 possible thermocyclers CFX96 (Bio-Rad) or (Agilent Technologies) - Confirmation: subculture of 0.1 ml BPW in 10 ml RVS broth, incubation 24 h ± 3 h at 41.5 C ± 1 C, streaking onto XLD or another chromogenic media, latex test (OXOID) or a biochemical gallery onto typical colonies without a purification step or by applying the tests described in the ISO 6579 standard after a purification step. Enrichment storage for 72 h at 2-8 C was also evaluated on positive samples in order to offer sufficient practicability to the users. This was done for the positive and discordant samples in the accuracy study. 2.2 Reference method The reference method corresponds to the ISO 6579 standard: Horizontal method for the detection of Salmonella spp. (See Appendix 2). 3 METHOD COMPARISON STUDY 3.1 Relative accuracy, relative specificity and relative sensitivity The relative accuracy is defined as the degree of correspondence between the response obtained by the reference method and the response obtained by the alternative method on identical samples. The relative specificity is the ability of the alternative method to not detect the analyte when it is not detected by the reference method. The relative sensitivity is the ability of the alternative method to detect the analyte when it is detected by the reference method. 3.1.1 Number and nature of samples 579 samples were analysed.the distribution per tested category and type is given in Table 1. Analyses performed according to the COFRAC accreditation ADRIA Développement 6/134 July 15, 2015

Table 1 Distribution per tested category and type Categories Meat and meat products Milks and dairy products Produces Eggs and egg products Fish and seafood products Feed products including pet food (25 g test portion) Pet foods (375 g test portion) Environmental samples Milk powders, infant formula with and without probiotics (375 g test portion) TOTAL Types Positive samples Negative samples Fresh and frozen meats 16 11 27 Ready-to-eat, ready-to-cook and ready-to-reheat meat preparations 12 11 23 Delicatessen (cured, fermented and cooked) 7 9 16 Total 35 31 66 Raw milks, fermented milks and cheeses 12 13 25 Milk powders, pasteurized milk and cheeses 11 10 21 Dairy desserts (Panna Cotta, ice cream, Tiramisu, etc.) 8 10 18 Total 31 33 64 Leafy greens and sprouts (spinach, lettuce, baby leaves, sprouts...) 9 19 28 RTE foods (composite deli-salads with fresh produces, fruit salads.) 12 6 18 Ready-to-cook and ready-to-reheat preparations 9 12 21 Total 30 37 67 Liquid eggs 10 9 19 Egg powders 10 9 19 RTE foods (Desserts, Pastries such custard based products, 10 12 22 M ayonnaises.) Total 30 30 60 Fish and seafood 10 10 20 Ready-to-eat, ready-to-cook and ready-to-reheat fish preparations 10 10 20 Smoked and cured fishes 10 10 20 Total 30 30 60 TOTAL INITIAL VALIDATION 156 161 317 Feed for pets (terrine, pâté, sausage) 13 9 22 Feed for cattle (flours, cereals, pellets) 8 14 22 Ingredients 12 10 22 Total 33 33 66 Ingredients 11 7 18 High moisture products (pâté, terrine, sausage) 9 11 20 Low moisture products (pelletts) 10 14 24 Total 30 32 62 Wipes, swabs, sponges 9 10 19 Dusts, wastes 7 11 18 Process and cleaning waters 15 15 30 Total 31 36 67 TOTAL EXTENSION VALIDATION n 1 94 101 195 Milks powders 11 12 23 Infant formula without probiotics 10 10 20 Infant formula with probiotics 12 12 24 Total Total 33 34 67 TOTAL EXTENSION VALIDATION n 2 33 34 67 283 296 579 ADRIA Développement 7/134 July 15, 2015

3.1.2 Artificial contamination of samples Artificial contaminations were done by spiking or seeding. For spiking, the strains were stressed using various injury protocols. The injury efficiency was evaluated by comparing enumeration results onto selective and non-selective agars (respectively XLD and TSYEA). The artificial contaminations are presented in Appendix 3. Initial validation study 153 samples were artificially contaminated, using 40 different strains. 136 samples gave a positive result. Most of the spiking inoculations, after injury protocols on the inoculum, were lower or equal to 5 CFU/sample. No more than 6 positive results were obtained with any single strain. First extension study (feed products, environmental samples, and pet food) 81 samples were artificially contaminated, using 29 different strains. 64 samples gave a positive result. No more than 6 positive results were obtained with any single strain. Second extension study (milk powders, infant formula with and without probiotics) 59 samples were artificially contaminated, using 15 different strains. 33 samples gave a positive result. No more than 6 positive results were obtained with any single strain. The repartition per contamination protocol and inoculation for all the categories is provided in Table 2 and Figures 1 and 2. ADRIA Développement 8/134 July 15, 2015

Table 2 - Repartition per contamination protocol and inoculation level for all the categories Spiking protocol Seeding protocol <5 cfu/sample 5<x<10 cfu/sample <3 cfu/sample 3< x < 7.2 cfu/sample Inoculated 186 35 50 22 293 Positive 155 34 25 19 233 Percentage of inoculation on positive samples Total 66.5 % 14.6 % 10.7 % 8.2 % 100 % Figure 1 - Ratio of positive samples after inoculation, depending on the inoculation protocol Seeding 3<x<7.2 cfu/sample Seeding <3 cfu/sample Spiking 5<x<10 cfu/sample 22 25 50 34 35 Positive samples Inoculated samples Spiking < 5 cfu/sample 155 186 0 50 100 150 200 Figure 2 - Breakdown of artificially inoculated samples, depending of the inoculation protocol 8,2% 14,6% 10,7% Spiking < 5 cfu/sample Spiking 5<x<10 cfu/sample Seeding <3 cfu/sample 66,5% Seeding 3<x<7.2 cfu/sample ADRIA Développement 9/134 July 15, 2015

The percentage of naturally and artificially contaminated samples is given Table 3. Table 3 Percentage of natural and artificial contaminated samples which provided positive results Initial validation Extension validation Extension validation Total study study n 1 study n 2 Number Number Number Number of positive Percentage of positive Percentage of positive Percentage of positive Percentage samples samples samples samples Artificial 136 87.2% 64 68.1% 33 100% 233 82.3 % Natural 20 12.8% 30 31.9% 0 0% 50 17.7% Total positive 156 / 94 / 33 / 283 / 17.7 % of the samples were naturally contaminated. 3.1.3 Confirmation protocols The positive PCR tests were confirmed by a subculture in RVS broth (0.1 ml + 10 ml) incubated for 24 h ± 3 h at 41.5 C prior to streaking onto XLD and a chromogenic media (ASAP). A latex test (OXOID Ref. FT0203A) and a biochemical gallery were then applied on typical colonies without a purification step. For unpaired data protocol (Dairy products), the whole protocol of the reference method was tested (RVS and MKTTn) in order to verify whether no Salmonella strain was detected in the enrichment broth. 3.1.4 Test results Raw data per category are given in Appendix 4. The paired results per category are given in the following tables, depending on the tested thermocyclers. ADRIA Développement 10/134 July 15, 2015

PP: positive presumptive non confirmed samples PD = positive deviation (R-/A+) ND = negative deviation (A-/R+) Table 4 Paired results of the reference method and alternative method: Overall results Thermocycler Response Reference method positive (R+) Reference method negative (R-) Alternative method Positive agreement (A+/R+) Positive deviation (R-/A+) CFX96 positive (A+) Alternative method PA = 265 Negative deviation (A-/R+) PD = 6 Negative agreement (A-/R-) negative (A-) ND = 12 NA = 296 (PPNA = 5) Alternative method Positive agreement (A+/R+) Positive deviation (R-/A+) positive (A+) Alternative method PA = 265 Negative deviation (A-/R+) PD = 5 Negative agreement (A-/R-) negative (A-) ND = 12 NA = 297 (PPNA = 10) Results per category of samples Table 5 Paired results of the reference method and alternative method: Meats and meat products Thermocycler Response Reference method positive (R+) Reference method negative (R-) Alternative method Positive agreement (A+/R+) Positive deviation (R-/A+) CFX96 positive (A+) Alternative method PA = 35 Negative deviation (A-/R+) PD = 0 Negative agreement (A-/R-) negative (A-) ND = 0 NA =31 Alternative method Positive agreement (A+/R+) Positive deviation (R-/A+) positive (A+) Alternative method PA = 35 Negative deviation (A-/R+) PD = 0 Negative agreement (A-/R-) negative (A-) ND = 0 NA = 31 (PPNA = 1) ADRIA Développement 11/134 July 15, 2015

Table 6 Paired results of the reference method and alternative method: Milks and dairy products Thermocycler Response Reference method positive (R+) Reference method negative (R-) Alternative method Positive agreement (A+/R+) Positive deviation (R-/A+) CFX96 positive (A+) Alternative method PA = 22 Negative deviation (A-/R+) PD = 6 Negative agreement (A-/R-) negative (A-) ND = 3 NA = 33 (PPNA = 1) Alternative method Positive agreement (A+/R+) Positive deviation (R-/A+) positive (A+) Alternative method PA = 21 Negative deviation (A-/R+) PD = 5 Negative agreement (A-/R-) negative (A-) ND = 4 NA = 34 (PPNA = 1) Table 7 Paired results of the reference method and alternative method: Produces Thermocycler Response Reference method positive (R+) Reference method negative (R-) Alternative method Positive agreement (A+/R+) Positive deviation (R-/A+) CFX96 positive (A+) Alternative method PA = 27 Negative deviation (A-/R+) PD =0 Negative agreement (A-/R-) negative (A-) ND = 3 NA = 37 (PPNA = 1) Alternative method Positive agreement (A+/R+) Positive deviation (R-/A+) positive (A+) Alternative method PA = 27 Negative deviation (A-/R+) PD = 0 Negative agreement (A-/R-) negative (A-) ND = 3 NA = 37 (PPNA= 3) Table 8 Paired results of the reference method and alternative method: Eggs and egg products Thermocycler Response Reference method positive (R+) Reference method negative (R-) Alternative method Positive agreement (A+/R+) Positive deviation (R-/A+) CFX96 positive (A+) Alternative method PA = 29 Negative deviation (A-/R+) PD = 0 Negative agreement (A-/R-) negative (A-) ND = 1 NA = 30 Alternative method Positive agreement (A+/R+) Positive deviation (R-/A+) positive (A+) Alternative method PA = 29 Negative deviation (A-/R+) PD = 0 Negative agreement (A-/R-) negative (A-) ND = 1 NA = 30 (PPNA = 1) ADRIA Développement 12/134 July 15, 2015

Table 9 Paired results of the reference method and alternative method: Fish and seafood products Thermocycler Response Reference method positive (R+) Reference method negative (R-) Alternative method Positive agreement (A+/R+) Positive deviation (R-/A+) CFX96 positive (A+) Alternative method PA = 29 Negative deviation (A-/R+) PD = 0 Negative agreement (A-/R-) negative (A-) ND = 1 NA = 30 Alternative method Positive agreement (A+/R+) Positive deviation (R-/A+) positive (A+) Alternative method PA = 29 Negative deviation (A-/R+) PD = 0 Negative agreement (A-/R-) negative (A-) ND = 1 NA = 30 Table 10 Paired results of the reference method and alternative method: Feed products, including pet foods (25 g test portion) Thermocycler Response Reference method positive (R+) Reference method negative (R-) Alternative method Positive agreement (A+/R+) Positive deviation (R-/A+) CFX96 positive (A+) Alternative method PA = 32 Negative deviation (A-/R+) PD = 0 Negative agreement (A-/R-) negative (A-) ND = 1 NA = 33 (PPNA = 1) Alternative method Positive agreement (A+/R+) Positive deviation (R-/A+) positive (A+) Alternative method PA = 32 Negative deviation (A-/R+) PD = 0 Negative agreement (A-/R-) negative (A-) ND = 1 NA = 33 Table 11 Paired results of the reference method and alternative method: Pet foods (375 g test portion) Thermocycler Response Reference method positive (R+) Reference method negative (R-) Alternative method Positive agreement (A+/R+) Positive deviation (R-/A+) CFX96 positive (A+) Alternative method PA = 28 Negative deviation (A-/R+) PD = 0 Negative agreement (A-/R-) negative (A-) ND = 2 NA = 32 Alternative method Positive agreement (A+/R+) Positive deviation (R-/A+) positive (A+) Alternative method PA = 29 Negative deviation (A-/R+) PD = 0 Negative agreement (A-/R-) negative (A-) ND = 1 NA = 32 (PPNA = 2) ADRIA Développement 13/134 July 15, 2015

Table 12 Paired results of the reference method and alternative method: Environmental samples Thermocycler Response Reference method positive (R+) Reference method negative (R-) Alternative method Positive agreement (A+/R+) Positive deviation (R-/A+) CFX96 positive (A+) Alternative method PA = 31 Negative deviation (A-/R+) PD = 0 Negative agreement (A-/R-) negative (A-) ND = 0 NA = 36 (PPNA = 1) Alternative method Positive agreement (A+/R+) Positive deviation (R-/A+) positive (A+) Alternative method PA = 31 Negative deviation (A-/R+) PD = 0 Negative agreement (A-/R-) negative (A-) ND = 0 NA = 36 (PPNA = 1) Table 13 Paired results of the reference method and alternative method: Milk powders, infant formula with and without probiotics (375 g test portion) Thermocycler Response Reference method positive (R+) Reference method negative (R-) Alternative method Positive agreement (A+/R+) Positive deviation (R-/A+) CFX96 positive (A+) Alternative method PA = 32 Negative deviation (A-/R+) PD = 0 Negative agreement (A-/R-) negative (A-) ND = 1 NA = 34 (PPNA = 1) Alternative method Positive agreement (A+/R+) Positive deviation (R-/A+) positive (A+) Alternative method PA = 32 Negative deviation (A-/R+) PD = 0 Negative agreement (A-/R-) negative (A-) ND = 1 NA = 34 (PPNA = 1) ADRIA Développement 14/134 July 15, 2015

3.1.5 Calculation of relative accuracy (AC), relative sensitivity (SE) and relative specificity (SP) (Appendix 5) Table 14 Calculation of relative accuracy (AC), relative sensitivity (SE) and relative specificity (SP) CFX96 thermocycler Relative Relative Relative Category PA NA ND PD N accuracy AC (%) N+ PA + ND sensitivity SE (%) N- NA + PD specificity SP (%) [100x(PA+NA])/N] [100xPA]/N+] [100xNA]/N-] Meat products 35 31 0 0 66 100.0 35 100.0 31 100.0 Dairy products 22 33 3 6 64 85.9 25 88.0 39 84.6 Vegetables 27 37 3 0 67 95.5 30 90.0 37 100.0 Egg products 29 30 1 0 60 98.3 30 96.7 30 100.0 Seafood products 29 30 1 0 60 98.3 30 96.7 30 100.0 Feed products including pet foods 32 33 1 0 66 98.5 33 97.0 33 100.0 (25 g test portion) Pet food (375 g test portion) 28 32 2 0 62 96.8 30 93.3 32 100.0 Environmental samples 31 36 0 0 67 100.0 31 100.0 36 100.0 Milk powders, infant formula with and without 32 34 1 0 67 98.5 33 97.0 34 100.0 probiotics (375 g test portion) All products 265 296 12 6 579 96.9 277 95.7 302 98.0 thermocycler Relative Relative Relative Category PA NA ND PD N accuracy AC (%) N+ PA + ND sensitivity SE (%) N- NA + PD specificity SP (%) [100x(PA+NA])/N] [100xPA]/N+] [100xNA]/N-] Meat products 35 31 0 0 66 100.0 35 100.0 31 100.0 Dairy products 21 34 4 5 64 85.9 25 84.0 39 87.2 Vegetables 27 37 3 0 67 95.5 30 90.0 37 100.0 Egg products 29 30 1 0 60 98.3 30 96.7 30 100.0 Seafood products 29 30 1 0 60 98.3 30 96.7 30 100.0 Feed products including pet foods 32 33 1 0 66 98.5 33 97.0 33 100.0 (25 g test portion) Pet food (375 g test portion) 29 32 1 0 62 98.4 30 96.7 32 100.0 Environmental samples 31 36 0 0 67 100.0 31 100.0 36 100.0 Milk powders, infant formula with and without 32 34 1 0 67 98.5 33 97.0 34 100.0 probiotics (375 g test portion) All products 265 297 12 5 579 97.1 277 95.7 302 98.3 ADRIA Développement 15/134 July 15, 2015

The alternative method percentage values are: Thermocycler CFX96 Relative accuracy : AC 96.9 97.1 Relative specificity : SP 98.0 98.7 Relative sensitivity : SE 95.7 95.7 Sensitivity of both methods, when the positive deviations of the alternative method are considered, is presented below: Thermocycler CFX96 Alternative method 95.8 95.7 Reference method 97.9 98.2 3.1.6 Analysis of discordants The discordant samples are the following, depending on the tested thermocycler. ADRIA Développement 16/134 July 15, 2015

Table 15 Negative deviations (CFX96: 12 ND : 12 ND) * DNA extract dilution: 1/4 Category Milks and dairy products Produces Eggs and egg products Fish and seafood products Feed products including pet foods Pet foods Milk powders, infant formula with and without probiotics Sample n Product CFX96 PCR result (ct) PCR result (ct) Confirmation 3732 Raw milk - - - 3739 Panna cotta i/+ (40.13)* i/-/-/- + 3740 Tiramisu - - - 4613 Pasteurised milk cheese 4476 Baby leaves 4485 Ready to reheat vegetables -/-/- -/-/- + -/-/+ (40.95) -/-/- -/-/- + -/+ (38.80/- 4572 Purée -/-/- -/-/- + 3691 White egg powder - - + 3574 Ready to reheat fish -/-/- -/-/- + 586 Cattle feed -/-/- -/-/- + 883 Raw material (poultry) 950 Pellets for cats 2626 Infant formula milk powder with probiotic -/-/+ (36.87) -/-/+ (39.01) + (37.32) + Natural -/-/+ (38.05) -/-/- -/-/- + TOTAL 12 12 + + Contamination (CFU/sample) Salmonella Montevideo Ad 912 (1.4) Salmonela Anatum Ad 298 (1.0) Salmonella Livingstone Ad 1169 (0.4) Salmonella Norwich Ad 1172 (1.0) Salmonella Panama Ad 1733 (3.0) Salmonella Orianenbourg Ad 1724 (1.0) Salmonella Virchow Ad 1721 (2.0) Salmonella Typhimurium 176 (1.2) Salmonella Indiana 2 (0.2) Salmonella Orianenburg Ad 1724 (2.6) Salmonella Livingstone F104 (3.2) Salmonella Cerro Ad 2152 (2.8) 12 negative deviations were observed with the CFX96 and the. For 10 samples, the confirmatory tests concluded to the presence of Salmonella in the enrichment broth. Note that for four samples (4476, 4485, 883 and 950), one PCR replicate was positive for the CFX96 cycler or the cycler. Indeed, the detection limit of the alternative method was probably not reached. For 2 samples (3732 and 3740), the presence of Salmonella spp. was not confirmed; the observed deviations are thus due to the sampling heterogeneity. ADRIA Développement 17/134 July 15, 2015

Table 16 Positive deviations (CFX96: 6 PD : 5 PD) Category Milks and dairy products Sample n Product CFX96 PCR result (ct) PCR result (ct) Contamination (CFU/sample) 3736 Fermented milk + (35.59) + (3519) Salmonella Montevideo Ad 912 (1.4) 3744 Milk powder + (38.71) -/+ (35.00) Salmonella Livingstone Ad 1169 (0.4) 3750 Raw milk cheese + (28.03) + (27.06) Salmonella Mbandaka Ad 1722 (2.2) 3789 Pasteurised milk cheese + (31.29) + (30.02) Salmonella Livingstone Ad 1169 (0.4) 3796 Dairy dessert + (20.88) + (19.25) Salmonella Ohio Ad 1482 (1.6) 4600 Raw milk + (31.00) + (31.25) Naturally contaminated Total 6 5 According to the Annex F of the ISO 16140 (2003), the analysis of discordant results, taking into account all the tested categories, is the following: Thermocycler CFX96 Y = ND + PD Y = 12 + 6 = 18 Y = 12 + 5 = 17 m 6 5 M 4 4 Conclusion m > M, the two methods are not different at < 0.05. Conclusion: according to the ISO 16140 (2003), the two methods are not different at < 0.05. 3.1.7 Confirmations The positive PCR tests were confirmed by BPW subculture in RVS broth and streaking onto XLD and a chromogenic agar plate (ASAP), followed by a latex test and a biochemical gallery on typical colonies. Typical colonies were observed with only one of both selective agar plates for 14 samples (See table 17). ADRIA Développement 18/134 July 15, 2015

Table 17 Sample n Inoculated strains Typical colonies XLD ASAP 3333 Naturally contaminated - + 3663 Salmonella Enteritidis MJG01 - + 4476 Salmonella Panama Ad 1733 - + 4477 Salmonella Panama Ad 1733 - + 4479 Salmonella Panama Ad 1733 - + 3749 Salmonella Mbandaka Ad 1722 - + 3750 Salmonella Mbandaka Ad 1722 - + 3751 Salmonella Mbandaka Ad 1722 - + 4607 Naturally contaminated - + 388 Naturally contaminated - + 805 Salmonella Agona A00V038 - + 283 Naturally contaminated - + 2429 Salmonella Dublin Ad 531 + - 2435 Salmonella Dublin Ad 531 + - All the typical colonies were confirmed by a latex test and a biochemical gallery without a purification step and the tests of the reference method. 3.1.8 BPW storage for 72 h at 2 8 C The enrichment broths were stored for 72 h at 2 8 C and analysed a second time. The following changes were observed: Sample Product Result before storage Result after storage n CFX96 CFX96 3315 Poultry meat PA PA PA ND 3739 Panna cotta PA ND ND ND 4476 Ready to reheat vegetables ND ND ND PA 3573 Ready to eat meal PA PA ND PA 582 Seeds PA PA ND PA 586 Cattle feed ND ND ND PA 883 Raw material ND PA PA PA 1873 Soya PA PA ND PA CFX96 Y = ND + PD Y = (12 + 3) + (6 + 0) = 21 Y = (12-1) + (5 + 0) = 16 m 6 5 M 5 3 Conclusion m > M, the two methods are not different at < 0.05. Conclusion: according to the ISO 16140 (2003), the two methods are not different at < 0.05. ADRIA Développement 19/134 July 15, 2015

3.2 Relative detection level For the purpose of this Standard, the relative detection level is the smallest number of culturable microorganisms than can be detected with 50 % of chances in the sample by the alternative and reference methods. 3.2.1 Matrices The objective of this study is (i) to determine the target species minimal quantity that can be detected in food matrices, (ii) to compare both method results. Detection limits were defined by analysing nine matrix/strain pairs. The tested matrices are given in Table 18. Table 18 Matrix Inoculated strain Storage Sample conditions size prior to analysis Ground beef Salmonella Typhimurium A00C060 48 h, 2 8 C 25 g Raw milk Salmonella Ohio Ad 1482 48 h, 2 8 C 25 g Spinach Salmonella Virchow Ad 1721 2 weeks, - 20 C 25 g Liquid egg product Salmonella Enteritidis Ad 638 48 h, 2 8 C 25 g Frozen fish fillet Salmonella Seftenberg Ad 355 2 weeks, - 20 C 25 g Pellets for dog Salmonella Mbandaka 2041 2 weeks, 20 C 25 g Pet food (pâté for dogs) Salmonella Derby Ad 1878 48 h, 2-8 C 375 g Process water Salmonella Livingstone A00E058 48 h, 2-8 C 25 g Milk powder with probiotics Salmonella Anatum Ad 298 2 weeks, 20 C 375 g 3.2.2 Contamination protocol Three contamination levels were tested: - 5 non-inoculated samples, - 20 samples inoculated at a low contamination level, providing fractional recovery, - 5 samples inoculated at a high contamination level. Three batches were prepared for each matrix: - One non-inoculated (150 g for a paired study, 300 g for an unpaired study), ADRIA Développement 20/134 July 15, 2015

- One inoculated between 0.5 and 1 CFU/25 g (750 g for a paired study, 1500 g for an unpaired study), - One inoculated at 2 CFU/25 g (150 g for a paired study, 300 g for an unpaired study). After inoculation, the different matrices were stored in the conditions described in Table 18 before being analysed by both methods. For 375 g sample size, 350 g of non-inoculated matrix was added in each stomacher bag prior to analysis». The background microflora was also enumerated. 3.2.3 Results Raw data are given in Appendix 6. Detection levels are presented in the Table 19, based on the ISO 16140 standard version (2003). Strain / matrix pairs Table 19 Relative detection level results Relative detection level (CFU / sample) according to Spearman-Kärber test 1 Alternative method Reference method (CFX96 and ) Ground beef / Salmonella Typhimurium A00C060 0.3 [0.2; 0.7] 0.3 [0.2; 0.7] Raw milk / Salmonella Ohio Ad 1482 0.4 [0.3; 0.6] 0.4 [0.3; 0.5] Spinach / Salmonella Virchow Ad 1721 0.3 [0.2; 0.5] 0.3 [0.2; 0.5] Liquid egg product / Salmonella Enteritidis Ad 638 0.2 [0.1; 0.4] 0.2 [0.1; 0.4] Frozen fish fillet / Salmonella Senftenberg Ad 355 0.4 [0.3; 0.6] 0.4 [0.3; 0.6] Pellets for dog / Salmonella Mbandaka Ad 2041 (25 g test portion) Pet food (pâté for dogs) / Salmonella Derby Ad 1878 (375 g test portion) 0.4 [0.3; 0.6] 0.4 [0.3; 0.6] 0.6 [0.4; 0.9] 0.6 [0.4; 0.9] Process water / Salmonella Livingstone A00E058 0.6 [0.4; 0.9] 0.6 [0.4; 0.9] Milk powder with probiotics / Salmonella Anatum Ad 298 0.6 [0.4; 0.9] 0.6 [0.4; 0.9] 3.2.4 Conclusion The relative detection levels of the two methods are similar; they are comprised between 0.1 and 0.9 CFU/sample (Spearman-Kärber test). 1 "Hitchins A. Proposed Use of a 50 % Limit of Detection Value in Defining Uncertainty Limits in the Validation of Presence-Absence Microbial Detection Methods, Draft 10th December, 2003". ADRIA Développement 21/134 July 15, 2015

3.3 Inclusivity / exclusivity Inclusivity is the ability of the alternative method to detect the target analyte from a wide range of strains. Exclusivity is the lack of interference from a relevant range of non-target strains of the alternative method. 3.3.1 Test protocols Inclusivity 100 Salmonella strains were cultured in BHI medium at 37 C. Dilutions were done in order to inoculate 10 cells/225 ml of BPW broth. The two enrichment protocols were then applied before performing the alternative method. Exclusivity 30 negative strains were cultured in BHI at 37 C. Dilutions were realized in order to inoculate 10 5 cells/ml BPW. BPW was incubated for 24 h at 37 C ± 1 C. The alternative method protocol was then used. 3.3.2 Results Raw data are given in Appendix 7. Inclusivity All 100 Salmonella strains gave a positive PCR result, for both enrichment protocols tested. Three strains gave negative latex tests when streaked onto XLD plates and seven when streaked onto ASAP. Exclusivity No cross reaction was observed among the 30 negative tested strains. 3.3.3 Conclusion The method is specific and selective. ADRIA Développement 22/134 July 15, 2015

3.4 Practicability The alternative method practicability was evaluated according to the AFNOR criteria relative to method comparison study. Packaging and reagents - For lysis: 1 x Lysis plate 1 x Domed caps 2 x Lysis buffer I 2 x Proteinase K - For PCR: 1 x 96 well PCR plate 1 x 96 optical caps 1 x yellow PCR support frame 2 x Positive Control DNA Storage conditions and shelf-life Specific equipment Reagents Training - Store 96 well PCR plates and Proteinase K at 20 C. - Store lysis buffer I at 4 C. Thaw Proteinase K just before adding it to lysis buffer - 2 x 96 well cooling block - 2 x Heating block 0.2 ml - Work rack S 96 systems (0.2 ml) Natural and blue - Capping / uncapping tool - Centrifuge for microtitter - Windows 7 PC with Microsoft Excel 2010 - One step (Dis) Assembling tool kit - Real-time PCR thermocycler - DNA Exitus Plus or 1 %-HCL - Nuclease free water One day is required for technicians with microbiology and molecular biology background. ADRIA Développement 23/134 July 15, 2015

Workflow (in minutes) for 24 samples Time to result (in days) Negative samples Steps Reference method Alternative method Sampling 48 48 Subculture in RVS 45 / Extraction and PCR / 30 BACGene evaluation sheet / / Streaking onto selective agar plates Reading selective agar plates Total for negative samples analyses 60 / 30 / 183 78 Total/negative sample 7.6 3.3 Presumptive samples or positive samples Steps Reference method Alternative method Subculture in RVS / 20 Streaking onto selective agar plates Reading selective agar plates / 10 / 10 test / 10 Confirmatory test 120 / Total for positive samples 303 128 Total/positive sample 12.6 5.3 Negative samples Steps Reference method Alternative method Sampling, pre-enrichment Day 0 Day 0 Subculture (RVS and MKTTn) Day 1 / Extraction and PCR / Day 1 Streaking onto selective agar plates Day 2 / Reading plates Day 3 / Presumptive positive or positive results Steps Reference method Alternative method Subculture in RVS / Day 1 Streaking onto selective agar plates / Day 2 Reading plates / Day 3 test / Day 3 Confirmatory tests Day 4 to Day 6 / ADRIA Développement 24/134 July 15, 2015

Technician background Technician qualified in microbiology and molecular biology Common step with the reference method Traceability of the results Maintenance Dairy products: no common step Other food products: sampling, pre-enrichment, subculture in RVS, streaking and reading selective agar plates All the data obtained with the are traced over time by the computer connected to the thermocycler using BACGene Evaluation Sheet Refer to the User Guide of the Thermocycler used. For technician assistance, contact EUROFINS GeneScan technical service or the local distributor. The method allows screening the negative samples within one day while 3 days are required with the reference method. 4 INTER-LABORATORY STUDY ORGANISATION AND RESULTS 4.1 Study organisation Samples were sent to 17 laboratories in February 2015. The study was done with vegetable mix inoculated by Salmonella Virchow Ad 1721 strain isolated from cereals. 24 samples were inoculated on Monday 16 th February 2015 for analysis by the reference method and the alternative method. An additional sample was added to the package for aerobic mesophilic flora enumeration by ISO 4833-1 method. The targeted inoculation levels were: - Level 0: 0 CFU/25 g, - Level 1: 2 CFU/25 g, - Level 2: 10 CFU/25 g. Blind coded samples were placed in isothermal boxes, which contained cooling blocks, and express-shipped to the different laboratories. ADRIA Développement 25/134 July 15, 2015

A temperature control flask containing a sensor was added to the package in order to register the temperature profile during the transport, the package delivery and storage until analyses. Samples were shipped in 24 h to the involved laboratories. The temperature conditions had to stay lower or equal to 8.4 C during transport, and between 0 C 8.4 C in the labs. Collaborators and ADRIA Développement carried out the analyses with the alternative and reference methods. 4.2 Experimental parameters control 4.2.1 Contamination level before inoculation, levels obtained after the artificial contaminations of the samples Before inoculation In order to detect Salmonella spp., the reference method was performed on five portions (25 g) before the inoculation. All the results were negative. Sample stability Sample stability was checked by inoculating the matrix at 100 CFU/g and 2 CFU/25 g. Enumerations were performed for the high contamination level and detection analyses were performed for the low contamination level. Triplicate samples were analysed, and the results were the following: Table 20 Day Reference method and CFU/g (XLD) alternative method (detection) Sample 1 Sample 2 Sample 3 Sample 1 Sample 2 Sample 3 Aerobic mesophilic flora (CFU/g) Day 0 + + + 89 55 73 1.1 10 4 Day 1 + + + 37 91 73 3.5 10 3 Day 2 + + + 100 73 100 4.4 10 4 No evolution was observed during storage at 2-8 C. ADRIA Développement 26/134 July 15, 2015

Contamination levels The contamination levels and the confidence intervals were: Table 21 - Contamination levels and confidence intervals Level Level 0 Low level High level Samples 3 5 7 11 16 19 22-24 1 2 6 9 10 13 17 20 4 8 12 14 15 18 21-23 Theoretical target level (b/25 g) True level (b/25 g sample) Low limit / 25 g sample High limit / 25 g sample 0 / / / 2 2.1 1.7 2.5 10 12.6 10.5 15.1 4.2.2 Logistic conditions The temperatures measured at reception by the Labs, the temperature registered by the thermo-probe, the receipt dates and analysis date are given in Table 22. Laboratories Temperature measured by the probe ( C) Table 22 - Sample temperatures at receipt Temperature measured at receipt ( C) Receipt date and time Analysis date and time A 1.5 2.5 17/02/2015 12h00 17/02/2015 20h15 B 1.5 4.5 17/02/2015 10h30 17/02/2015 14h00 C 2.5 4.0 17/02/2015 12h00 17/02/2015 13h00 D 0.5 2.9 17/02/2015 11h15 17/02/2015 14h00 E Not received 9.8 17/02/2015 12h00 18/02/2015 16h20 F 0.5 3.9 17/02/2015 12h00 17/02/2015 13h00 G 1.0 2.0 17/02/2015 11h10 17/02/2015 13h15 H 1.0 2.8 17/02/2015 11h30 17/02/2015 12h00 I 1.0 2.7 17/02/2015 11h00 17/02/2015 15h00 J 0.0 2.0 17/02/2015 09h00 17/02/2015 14h00 K 1.0 4.7 17/02/2015 09h30 17/02/2015 13h00 L 3.5 3.7 17/02/2015 09h15 17/02/2015 09h15 M 1.5 4.0 17/02/2015 13h00 18/02/2015 15h00 N 3.0 7.4 17/02/2015 19h50 18/02/2015 11h30 O 0.5 1.5 17/02/2015 08h30 18/02/2015 14h00 P 1.0 2.3 17/02/2015 09h46 17/02/2015 13h00 Q 1.5 4.7 17/02/2015 10h30 17/02/2015 12h00 ADRIA Développement 27/134 July 15, 2015

4.2.3 Conclusion All the packages were delivered at Day 1 (17/02/2015). No problems were encountered during shipment. One Lab measured a temperature receipt above the limit (Lab E: 9.8 C). We have not yet received the probe, but we have considered, regards to the other temperatures measured at receipt, that the temperature was also correct for this Lab. 4.3 Results analysis 4.3.1 Aerobic mesophilic flora enumeration For all the Labs (except Lab E), the enumeration of the aerobic mesophilic flora ranged between 4.2 10 3 to 2.9 10 4 CFU/g. Lab E got an enumeration value of 9.6 10 6 CFU/g as they incubated the suspension before enumeration. 4.3.2 Expert lab results The raw data are given in Appendix 8. They are summarised in Table 23. Table 23 Results obtained by the expert Lab. Level Reference method Alternative method L0 0/8 0/8 L1 8/8 8/8 L2 8/8 8/8 4.3.3 Collaborator lab results 17 Labs participated to the study. 13 Labs realised the analyses at Day 1, and 4 at Day 2 (Labs E, M, N and O). All the raw data are given in Appendix 9. The results are summarised in Table 24. ADRIA Développement 28/134 July 15, 2015

Table 24 Results obtained by the collaborator Labs. Reference method Laboratory L0 L1 L2 A 0/8 8/8 8/8 B 0/8 7/8 8/8 C 0/8 7/8 8/8 D 0/8 7/8 8/8 E 0/8 7/8 8/8 F 0/8 8/8 8/8 G 0/8 8/8 8/8 H 0/8 8/8 8/8 I 0/8 7/8 8/8 J 4/8 8/8 8/8 K 0/8 8/8 8/8 L 0/8 8/8 8/8 M 3/8 7/8 8/8 N 0/8 8/8 8/8 O 0/8 7/8 8/8 P 0/8 7/8 8/8 Q 0/8 7/8 8/8 Alternative method (before confirmation) Alternative method (after confirmation) Laboratory L0 L1 L2 Laboratory L0 L1 L2 A 0/8 8/8 8/8 A 0/8 8/8 8/8 B 0/8 7/8 8/8 B 0/8 7/8 8/8 C 0/8 7/8 8/8 C 0/8 7/8 8/8 D 0/8 7/8 8/8 D 0/8 7/8 8/8 E 0/8 7/8 8/8 E 0/8 7/8 8/8 F 0/8 8/8 8/8 F 0/8 8/8 8/8 G 0/8 8/8 8/8 G 0/8 8/8 8/8 H 0/8 8/8 8/8 H 0/8 8/8 8/8 I 0/8 7/8 8/8 I 0/8 7/8 8/8 J 0/8 8/8 7/8 J 0/8 8/8 7/8 K 0/8 8/8 8/8 K 0/8 8/8 8/8 L 0/8 8/8 8/8 L 0/8 8/8 8/8 M 3/8 7/8 8/8 M 0/8 7/8 8/8 N 0/8 8/8 8/8 N 0/8 8/8 8/8 O 0/8 7/8 8/8 O 0/8 7/8 8/8 P 0/8 8/8 8/8 P 0/8 7/8 8/8 Q 0/8 7/8 8/8 Q 0/8 7/8 8/8 ADRIA Développement 29/134 July 15, 2015

Two Labs (J and M) obtained positive results on control samples: - Lab J: 2 for the reference method - Lab M: 1 for the reference method 1 for the alternative method (PCR test). Nine Labs observed one negative result with the reference and the alternative methods for the low contamination level (B, C, D, E, I, M, O, P and Q). Two Labs obtained one positive PCR test (M and P) which were not confirmed. Note that for Lab P, a late Ct value was obtained (40.04); this was probably due to a cross contamination. Finally, 15 Labs were kept for the interpretation; the results of the two Labs (J and M) which had positive control samples were not taken into account. Among the 15 Labs retained, 12 used the and 3 the CFX96. 4.4 Results interpretation 4.4.1 Specificity and sensitivity for each method For the L0 level and for each method, specificity percentages are calculated according to: FP SP 1 x 100% N with: N- = total number of all L0 assays FP = number of false positive results For each contamination level and each method, the sensitivity percentages are calculated according to: TP SE x 100% N with : N+ = total number of all L1 or L2 assays TP = number of true positive results ADRIA Développement 30/134 July 15, 2015

Results (See Appendix 10) are reported in Table 25. Table 25 Interpretation Level Reference method Alternative method SP/SE % LCL% SP/SE % LCL% L0 (SP) 100.0 98.00 100.0 98.0 L1 (SE) 93.3 88.3 93.3 88.8 L2 (SE) 100.0 98.0 100.0 98.0 L1+L2 (SE) 96.7 93.4 96.7 93.4 LCL: confidence interval 4.4.2 Relative accuracy (AC) Results for all levels (See Appendix 11) are given below: Table 26 - Paired results of the alternative and reference methods Alternative method Reference method + - Total + PA = 232 PD = 0 232 - ND = 0 NA = 128 (PPNA = 1) 128 Total N+ = 232 N- = 128 N = 360 Relative accuracy (AC) (in %) is calculated according to: ( PA NA) AC x 100% N with : N = number of samples analysed PA = number of positive agreement NA = number of negative agreement The alternative method accuracy values with regard to the reference method are: Table 27 Interpretation Level AC % LCL % L0 100.0 98.0 L1 100.0 98.0 L2 100.0 98.0 L1 + L2 100.0 98.0 Total 100.0 98.0 ADRIA Développement 31/134 July 15, 2015

4.4.3 Discordant results No discordant result was observed; the two methods are equivalent. 4.5 Interpretation 4.5.1 Comparison of the relative accuracy, specificity and sensitivity values The values obtained for the two parts of the validation study (comparative and inter-laboratory studies) are reported in Table 28. Table 28 - Alternative method values calculated during the comparative and inter-laboratory studies Inter-laboratory study Method comparative study CFX96 Relative accuracy (AC) 100 % 95.6 % 95.6 % Sensitivity (SE) 96.7 % 94.7 % 94.0 % Specificity (SP) 100.0 % 96.4 % 97.0 % 4.5.2 Accordance (DA) Accordance values for both methods are (See Appendix 12): Table 29 Interpretation Level Reference method (DA) Alternative method (DA) L0 100.0 % 100.0 % L1 88.3 % 88.3 % L2 100.0 % 100.0 % ADRIA Développement 32/134 July 15, 2015

4.5.3 Concordance Both methods concordance values are (See Appendix 13): Table 30 Interpretation Level Reference method Alternative method L0 100.0 % 100.0 % L1 87.5 % 87.5 % L2 100.0 % 100.0 % 4.5.4 Odds Ratio (COR) The odds ratio value is determined according to: Both method odds ratio values are: Accordance x 100 condorcance COR Concordance x (100 accordance ) Table 31 Interpretation Level Reference method (COR) Alternative method (COR) L0 1.00 1.00 L1 1.08 1.08 L2 1.00 1.00 ADRIA Développement 33/134 July 15, 2015

5 CONCLUSION The methods comparative study conclusions are: The validation study was run with 2 thermocyclers: CFX96 from Bio-Rad or from Agilent Technologies. The method shows: - A satisfying sensitivity study, - Limit of detections that are similar to the ISO 6579 standard. The method is specific and selective. It is possible to store the enrichment broth for 72 h at 2-8 C prior running the PCR test. The inter-laboratory study conclusions are: The observed data and results confirmed that the alternative method and reference method show equivalent performances (accordance, concordance, odds ratio). ADRIA Développement 34/134 July 15, 2015

Appendix 1 Flow diagrams of the alternative method All food products except dairy products Dairy products 25 g + 225 ml BPW* 20 h ± 4 h at 37 C ± 1 C 25 g + 225 ml BPW* 21 h ± 3 h at 41.5 C ± 1 C By using serological pipette: remove an aliquot of enrichment culture and dispense into a sterile container Prepare lysis buffer by adding proteinase K Dispense 90 µl into lysis tubes Add 10 µl of enrichment sample to lysis buffer (100 µl total) Incubate at 37 C for 20 min Incubate at 95 C for 10 min Remove from heat Cool for 5 min Briefly centrifuge samples Transfer 5 µl of lysate to PCR tubes Run PCR. Approx. 1 h 40 min Confirmation: Subculture in RVS broth (0.1 ml + 10 ml) for 24 h ± 3 h at 41.5 C ± 1 C Streaking onto selective agar plates (XLD or a chromogenic agar) tests on typical colonies without a purification step or biochemical galleries or Subculture on a non-selective agar plate Tests described in the ISO 6579 standard * Sample preparation according to the ISO 6887 parts ADRIA Développement 35/134 July 15, 2015

Feed products including pet foods (25 g test portion) 25 g + 225 ml BPW Pet foods (375 g test portion) 375 g + 3 375 ml pre-warmed BPW Environmental samples Swab + 10 ml BPW Sponge or wipe + 225 ml BPW Wastes, dusts, process water: 25 g + 225 ml BPW Incubation 21 h ± 3 h at 37 C ± 1 C By using serological pipette: remove an aliquot of enrichment culture and dispense into a sterile container Prepare lysis buffer by adding proteinase K Dispense 90 µl into lysis tubes Add 10 µl of enrichment sample to lysis buffer (100 µl total) Incubate at 37 C for 20 min Incubate at 95 C for 10 min Remove from heat Cool for 5 min Briefly centrifuge samples Transfer 5 µl of lysate to PCR tubes Run PCR. Approx. 1 h 40 min Confirmation: Subculture in RVS broth (0.1 ml + 10 ml) for 24 h ± 3 h at 41.5 C ± 1 C Streaking onto selective agar plates (XLD or a chromogenic agar) tests on typical colonies without a purification step or biochemical galleries or Subculture on a non selective agar plate Tests described in the ISO 6579 standard ADRIA Développement 36/134 July 15, 2015

Milk powder without probiotics 375 g * + 3 375 ml pre-warmed BPW Milk powder with probiotics 375 g * + 3 375 ml double strength pre-warmed BPW Incubation 21 h ± 3 h at 37 C ± 1 C By using serological pipette: remove an aliquot of enrichment culture and dispense into a sterile container Prepare lysis buffer by adding proteinase K Dispense 90 µl into lysis tubes Add 10 µl of enrichment sample to lysis buffer (100 µl total) Incubate at 37 C for 20 min Incubate at 95 C for 10 min Remove from heat Cool for 5 min Briefly centrifuge samples Transfer 5 µl of lysate to PCR tubes Run PCR. Approx. 1 h 40 min Confirmation: Subculture in RVS broth (0.1 ml + 10 ml) for 24 h ± 3 h at 41.5 C ± 1 C Streaking onto selective agar plates (XLD or a chromogenic agar) tests on typical colonies without a purification step or biochemical galleries or Subculture on a non selective agar plate Tests described in the ISO 6579 standard * Sample preparation according to the ISO 6887 parts ADRIA Développement 37/134 July 15, 2015

Appendix 2 ISO 6579:2002: Microbiology of food and animal feeding stuffs - Horizontal method for the detection of Salmonella spp. 25g of sample + 225 ml BPW Incubation 18 h 2 h at 37 C 1 C 0.1 ml BPW 1 ml BPW 10 ml RVS 10 ml MKTTn Incubation 24 h 3 h at 41.5 C 1 C Incubation 24 h 3 h at 37 C 1 C Streak on XLD and a second medium Incubation 24 h 3 h at 37 C 1 C Streak 1 characteristic colony onto nutrient agar (Take 4 other colonies if the first one is negative) Incubation 24 h 3 h at 37 C 1 C Biochemical and serological confirmation ADRIA Développement 38/134 July 15, 2015

Milk powder without probiotics 375 g * + 3 375 ml pre-warmed BPW Milk powder with probiotics 375 g * + 3 375 ml double strength pre-warmed BPW Incubation 18 h 2 h at 37 C 1 C 0.1 ml BPW 1 ml BPW 10 ml RVS 10 ml MKTTn Incubation 24 h 3 h at 41.5 C 1 C Incubation 24 h 3 h at 37 C 1 C Streak on XLD and ASAP Incubation 24 h 3 h at 37 C 1 C Streak 1 characteristic colony onto nutrient agar (Take 4 other colonies if the first one is negative) Incubation 24 h 3 h at 37 C 1 C Biochemical and serological confirmation * Sample preparation according to the ISO 6887 parts ADRIA Développement 39/134 July 15, 2015

Appendix 3 Artificial contamination of the samples N Sample Product Artificial contaminations (spiking protocol) Strain Origin Injury Injury Inoculation protocol measurement level/25 g 4568 Fresh vegetables Salmonella Agona Ad1725 Cereals HT 56 c 8 min 0.39 3-7-2-4-5 (4.2) + 4569 Fresh vegetables Salmonella Agona Ad1725 Cereals HT 56 c 8 min 0.39 3-7-2-4-5 (4.2) + 4609 Pasteurized milk Salmonella Anatum Ad1168 Dairy product HT 56 Cc 8 min 1.48 1-1-2-3-2 (1.8) + 4611 Pasteurized milk cheese Salmonella Anatum Ad1168 Dairy product HT 56 Cc 8 min 1.48 1-1-2-3-2 (1.8) + 3664 Raw fish fillet Salmonella Anatum Ad1451 Fish fillet 12 days 4 C 0.47 5-8-9-5-4 (6.2) + 3665 Raw fish fillet Salmonella Anatum Ad1451 Fish fillet 12 days 4 C 0.47 5-8-9-5-4 (6.2) + 3666 Raw haddock fillet Salmonella Anatum Ad1451 Fish fillet 12 days 4 C 0.47 5-8-9-5-4 (6.2) + 3667 Raw coley filet Salmonella Anatum Ad1451 Fish fillet 12 days 4 C 0.47 5-8-9-5-4 (6.2) + 3695 Raw fish Salmonella Anatum Ad1451 Fish fillet 18 days 4 C 0.35 0-6-2-4-2 (2.8) + 3697 Salted cod Salmonella Anatum Ad1451 Fish fillet 18 days 4 C 0.35 0-6-2-4-2 (2.8) + 3738 Dairy based dessert(panna cotta) Salmonella Anatum Ad298 Milk powder HT 56 C 8min 1.81 3-1-0-0-1 (1.0) + 3739 Dairy based dessert(panna cotta) Salmonella Anatum Ad298 Milk powder HT 56 C 8min 1.81 3-1-0-0-1 (1.0) + 3743 Skimmed milk powder Salmonella Anatum Ad298 Milk powder HT 56 C 8min 1.81 3-1-0-0-1 (1.0) + 3795 Milky dessert Salmonella Anatum Ad298 Milk powder HT 56 C 8min/6 days 4 C 1.12 4-4-0-1-1 (2.0) - 3797 Milk powder with probiotic Salmonella Anatum Ad298 Milk powder HT 56 C 8min/6 days 4 C 1.12 4-4-0-1-1 (2.0) - 3798 Milk powder with probiotic Salmonella Anatum Ad298 Milk powder HT 56 C 8min/6 days 4 C 1.12 4-4-0-1-1 (2.0) - 4488 Smoked haddock Salmonella Brandenburg Ad1351 Fish 8 days 4 C 0.68 3-0-3-4-3 (3.2) + 4489 Smoked mackerel Salmonella Brandenburg Ad1351 Fish 8 days 4 C 0.68 3-0-3-4-3 (3.2) + 3378 RTRH meal(chicken and purée) Salmonella Bredeney Ad2042 Turkey meat HT 56 C 8min 1.24 2-3-2-2-1 (2.0) + 3380 RTRH meal(chicken and rice) Salmonella Bredeney Ad2042 Turkey meat HT 56 C 8min 1.24 2-3-2-2-1 (2.0) + 3381 RTRH meal(chicken ) Salmonella Bredeney Ad2042 Turkey meat HT 56 C 8min 1.24 2-3-2-2-1 (2.0) + 3383 RTRH meal(chicken and rice) Salmonella Bredeney Ad2042 Turkey meat HT 56 C 8min 1.24 2-3-2-2-1 (2.0) + 3385 RTRH meal(chicken, mushrooms and rice) Salmonella Bredeney Ad2042 Turkey meat HT 56 C 8min 1.24 2-3-2-2-1 (2.0) + 3289 Raw fish Salmonella Derby Ad1093 Fish fillet 5 days -20 C 0.63 0-1-2-2-5 (2.0) + 3290 Raw cod Salmonella Derby Ad1093 Fish fillet 5 days -20 C 0.63 0-1-2-2-5 (2.0) + 3291 Salmon Salmonella Derby Ad1093 Fish fillet 5 days -20 C 0.63 0-1-2-2-5 (2.0) + 3457 Cooked beef Salmonella Derby Ad1879 Meat 6 days 10% NaCl 0.48 4-0-4-1-7 (3.2) + 4490 salted cod Salmonella Derby F81 Fish 8 days 10% NaCl 0.54 1-2-0-1-2 (1.2) - 3297 Egg white powder Salmonella Enteritidis 10 Egg white powder HT 56 C 8min 0.44 3-4-6-4-6 (4.6) + 3298 Egg yolk powder Salmonella Enteritidis 10 Egg white powder HT 56 C 8min 0.44 3-4-6-4-6 (4.6) + 3299 Whole egg powder Salmonella Enteritidis 10 Egg white powder HT 56 C 8min 0.44 3-4-6-4-6 (4.6) + 3586 Egg based dessert Salmonella Enteritidis 465 Liquid egg product HT 56 C 8min 1.84 5-7-5-4-9 (6.0) + ADRIA Développement 40/134 July 15, 2015 Global result