Quantitative Real-Time PCR Detection of Inactivated H5 Avian Influenza Virus in Raw Milk Samples by Miniaturized Instruments Designed for On-Site Testing

16 Highly pathogenic avian influenza virus (HPAIV) of H5 and H7 subtypes has emerged 17 as one of the most important zoonotic pathogens in the 21 st century with significant economic 18 consequences. The recent outbreak of H5N1 avian influenza (AI) in dairy cattle highlighted 19 the importance of early detection in managing and mitigating HPAIV outbreaks. A successful 20 high-speed diagnostic response requires rapid site and specimen access, minimal time for test 21 protocols, and prompt communication of the diagnostic results to government officials. A new 22 diagnostic paradigm that consists of miniaturized extractor and qPCR instruments 23 (EZextractor and EZcycler MiniQ), designed for mobile, on-site testing has been compared 24 with a platform of benchtop instruments (QIAGEN RNeasy and QuantStudio™ 5) for 25 detecting inactivated H5 avian influenza virus (AIV) spiked in raw milk samples. Two sets of 26 experiments were performed: 1) 15 raw milk samples, obtained from 15 different farms, 27 diluted with phosphate-buffered saline and spiked with the virus to reach approximately 10 28 copies/µL virus concentration, and 2) raw milk samples from two farms, each spiked with the 29 inactivated AIV H5 followed by 5 series of dilution to reach AIV concentrations of 1000, 100, 30 10, 1 and 0.1 copies/µL. Results show that despite the inhibitors in raw milk, AIV in all 31 samples can be detected by both platforms. The MT platform showed higher sensitivity than 32 the benchtop platform: the Ct values from the MT were ~2 units lower than the benchtop Ct 33 values. Our findings demonstrate the robustness of the MT platform for diagnosing AIV H5 in 34 raw milk samples and support its use as an on-site diagnostic for rapid surveillance and 35 response. 36 37

Avian influenza, H5 avian influenza virus, molecular testing, qPCR, pathogen 38 .CC-BY-NC-ND 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted June 5, 2025. ; https://doi.org/10.1101/2025.06.02.657307doi: bioRxiv preprint 2 detection, avian influenza virus in raw milk, on-site testing 39

40 Avian influenza virus (AIV) has been a significant threat to the poultry industries 41 worldwide, with global outbreaks caused primarily by the AIV of H5 and H7 subtypes 42 since the mid-20 th century [1]. The current wave of avian influenza started in 2020 43 and is dominated by the highly pathogenic avian influenza virus (HPAIV) H5N1 44 strain of the clade 2.3.4.4b, which originated in Europe and first spread to North 45 America in 2021 with frequent spillover events to mammals reported [1, 2]. In the 46 United States, the potential spread of the HPAIV H5N1 strain beyond poultry had a 47 significant upturn when HPAIV H5N1 was isolated from milk in dairy cattle in Texas 48 in February 2024 with cattle showing symptoms such as reduced milk production and 49 decreased feed intake [3]. Zoonotic infection of HPAIV H5N1 of dairy farm workers 50 has also been reported, although the symptoms were milder compared to 51 bird-transmitted HPAIV H5N1 infection [4, 5]. 52 53 The implications in human public health have been further highlighted by the 54 fact that 70 confirmed zoonotic HPAIV H5 human cases have been reported in the 55 United States as of May 2025 [6,7]. A new genotype D1 typically carried by wild 56 birds was found to be responsible for the first severe HPAIV H5N1 case in Canada in 57 November 2024, and the first severe case in the United States in Louisiana in 58 December 2024 that has resulted in death [8,9]. 59 60 Since its first discovery in cattle milk, the HPAIV H5N1 has since spread in the 61 United States and as of May 2025, at least 1,052 dairy herds across 17 states have 62 been affected, prompting urgent governmental action to address the growing crisis [6]. 63 California, the largest milk-producing state in the United States bears the brunt of this 64 outbreak accounting for 73% of cases reported in the United States [6]. In response, 65 the state government of California has declared a state of emergency to mobilize 66 resources to combat the virus's rapid spread and uphold consumer confidence in dairy 67 products [10]. Nationwide, to address the importance of early detection and 68 containment to mitigate the spread of H5N1 within dairy herds and protect public 69 health, the United States Department of Agriculture (USDA) announced a new 70 Federal Order to implement a National Milk Testing Strategy on December 6, 2024 71 [11]. Per the Order, milk samples were to be tested by molecular testing using 72 quantitative reverse transcription polymerase chain reaction (qRT-PCR), which is 73 considered the standard used in the detection of AIV according to the USDA ’s Animal 74 and Plant Health Inspection Service [12]. 75 76 .CC-BY-NC-ND 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted June 5, 2025. ; https://doi.org/10.1101/2025.06.02.657307doi: bioRxiv preprint 3 The implementation of the Federal Order for milk testing will challenge the 77 current capacity of qRT-PCR diagnostic testing in many state diagnostic laboratories. 78 Further, the current laboratory diagnostic routine from taking samples to 79 transportation to a laboratory to results takes time that is critical for tracing the 80 original source of infection. Here we present a new mode of qRT-PCR diagnosis using 81 miniaturized extraction and PCR instruments that can be housed in a mobile 82 laboratory for conducting the diagnosis on-site. The mobile testing (MT) platform for 83 detecting H5N1 in milk samples offers a critical solution for rapid surveillance and 84 early intervention. By enabling on-site and timely detection, MT platform reduces the 85 lag between sample collection and diagnosis, enhancing the ability to contain 86 outbreaks before they escalate. 87 88 This study aimed to evaluate the effectiveness of the MT platform for detecting 89 the AIV H5 in raw milk samples and its adaptability to different operational settings 90 by comparing different PCR reagents and systems. 91 92 93

94 Virus. Three samples of lysis buffer-inactivated AIV H5 (clade 2.3.4.4b) taken from a 95 goose autopsy tissue/organ sample in Pingtung County, Taiwan, were obtained as part 96 of a research grant awarded to Ya-Mei Chen from the Animal and Plant Health 97 Inspection Agency, Taiwan Ministry of Agriculture (research grant: Research and 98 Analysis of Key Avian Disease Surveillance and Control, ID: 113AS-5.5.5-VP-01; In 99 Chinese: 家禽重要疾病監測及防控研析 , ID: 113 農科 -5.5.5- 檢 -01). The 100 concentrations of inactivated AIV H5 suspended in the lysis buffer of the three 101 samples are estimated based on quantitative PCR measurements using the positive 102 control of the ThermoFisher AIV kit with a known RNA concentration of 1,000 103 copies/µL. The estimated concentrations of samples 1, 2, and 3 are 1,000, 10,000, and 104 1,000 copies/µL, respectively. All experiments were conducted in the laboratories of 105 the Schweitzer Biotech Company in Taipei, Taiwan. 106 107 Milk Samples. Fifteen fresh raw milk samples were collected from different dairy 108 farms in Taiwan. All milk samples were diluted with phosphate-buffered saline (PBS) 109 at a ratio of 1:3 by volume to reduce inhibitory effects of components in the milk. The 110 fifteen raw milk samples were used for evaluating the performance of MT platform 111 (EZextractor + MiniQ) and the USDA-recommended benchtop platform (RNeasy + 112 QS5) in detecting AIV H5. Each of the 15 raw milk samples was spiked with either of 113 .CC-BY-NC-ND 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted June 5, 2025. ; https://doi.org/10.1101/2025.06.02.657307doi: bioRxiv preprint 4 the three inactivated AIV H5 samples (Samples 1, 2, or 3) and diluted to reach a 114 concentration of 10 copies/µL. The spiking and dilution process is outlined in Table 1. 115 116 Sample 1 Sample 2 Sample 3 Stock concentration 1,000 copies/µL 10,000 copies/µL 1,000 copies/µL Diluent Milk-PBS mixture (1:3) Milk-PBS mixture (1:3) Milk-PBS mixture (1:3) Dilution method* Two ten-fold dilutions Three ten-fold dilutions Two ten-fold dilutions Target concentration 10 copies/µL 10 copies/µL 10 copies/µL Table 1. Inactivated AIV H5 in milk samples used in this study. 117 *First dilution was made by adding 40 µL of inactivated virus sample to 100 µL of milk and 118 260 µL PBS to achieve a concentration of 25% milk by volume. Subsequent ten-fold dilutions 119 were made by adding 100 µL of the prior sample solution to 900 µL of diluent. 120 121 Polymerase Chain Reaction. The EZextractor Viral DNA/RNA Extraction Kit (Ref: 122 ATXZ008, Schweitzer Biotech Company, Taipei, Taiwan) was used to isolate viral 123 RNA using the EZextractor Nucleic Acid Extraction System (Model M32, Schweitzer 124 Biotech Company, Taipei, Taiwan) following manufacturer’s built-in program V9. 125 The EZextractor system is a fully automated magnetic bead-based platform designed 126 to streamline nucleic acid extraction for laboratories handling medium or large sample 127 sizes. It operates within a compact footprint and features a touchscreen interface for 128 user control. The system supports flexible batch processing with capacities of up to 32 129 samples per run with an approximate runtime of 30 minutes. As a comparison in this 130 study, QIAGEN RNeasy Mini kit (Cat. #74106 250, QIAGEN N.V ., Venlo, The 131 Netherlands) was used to extract AIV H5 RNA spiked into raw milk samples 132 following the Protocol: Purification of Total RNA from Animal Cells Using Spin 133 Technology according to the kit handbook. 134 135 Extracted AIV RNA was then amplified by VetMAX™ Gold AIV Detection Kit 136 (Cat# 4485261, ThermoFisher Scientific, Waltham, MA) using Influenza Virus Primer 137 Probe Mix included in the kit (FAM reporter dye). The following control samples 138 were also included in the kit: Influenza Virus-Xeno™ RNA Control Mix for positive 139 control; Xeno™ RNA Control for internal positive control (VIC reporter dye). 140 141 The qRT-PCR process was performed and analyzed using two types of qPCR 142 thermocyclers: the benchtop QuantStudio™ 5 (QS5) Real-Time PCR System (Cat# 143 .CC-BY-NC-ND 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted June 5, 2025. ; https://doi.org/10.1101/2025.06.02.657307doi: bioRxiv preprint 5 A28133, ThermoFisher Scientific, Waltham, MA) and the portable EZcycler Mini 144 real-time PCR (MiniQ) System (Schweitzer Biotech Company, Taipei, Taiwan). The 145 MiniQ PCR System is a compact, space-efficient system for precise quantitative 146 analysis suitable for MT. Equipped with a built-in touchscreen, it operates 147 independently of a computer and has a sample capacity of 16 and three detection 148 channels. Two independent qPCR runs were performed for each of the milk samples. 149 150 The thermal cycling conditions in reverse transcription followed by qPCR 151 amplification for both thermocyclers were as follows: 152 153 Stage Cycles. Temperature Time Reverse transcription 1 48°C 10 minutes RT inactivation/initial denaturation 1 95°C 10 minutes Amplification: Denaturation Amplification: Annealing and extension 40 95°C 60°C 15 seconds 45 seconds* Table 2. Reverse transcription and qPCR thermocycling conditions. 154 *: Fluorescence signal is collected at this step. 155 156 Sensitivity of AIV H5 Detection. Serial ten-fold dilutions of the inactivated virus 157 sample 2 stock solution (10,000 copies/µL) using raw milk-PBS mixture (raw milk 158 from farm 8 or farm 9) were performed to determine the limits of detection and the 159 efficiency of the MT platform. 160 161 Raw Milk Inhibitors. To assess if the potential inhibitors in raw milk influence the 162 detection of AIV , PBS solutions were spiked with inactivated AIV H5. Two raw milk 163 samples were randomly chosen from the stocks and diluted with PBS at 1:3 ratio. 164 Each sample, either raw milk in PBS or PBS only, was spiked using the same source 165 with an unspecified quantity of inactivated viruses. Internal positive control was also 166 used to spike the PBS only as well as the raw milk diluted with PBS. In this 167 comparison, one set of experiments was performed using the combination of 168 Ezextractor and the benchtop QS5, and the other Ezextractor and the miniaturized 169 MiniQ. 170 171 172

173 Both platforms successfully detected AIV H5 in all 15 spiked raw milk samples 174 .CC-BY-NC-ND 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted June 5, 2025. ; https://doi.org/10.1101/2025.06.02.657307doi: bioRxiv preprint 6 at 10 copies/µL viral concentration (Table 3 and Figure 1). Spearman’s correlation of 175 these two assays is 0.96, which shows good consistency between the MT platform and 176 the benchtop platform. The paired t test shows that there is a significant difference in 177 the Ct values between these two assays: The MT assay’s Ct values are 2 units smaller 178 than that of the USDA benchtop assay. These results validate the performance of the 179 MT platform for detecting AIV in raw milk. 180 181 182 Table 3. Ct values of qPCR run of 15 raw milk samples spiked with inactivated AIV H5 using 183 the MT and the USDA benchtop platforms. Two independent PCR runs were performed for 184 each sample using each platform. PC: positive control. NTC: no template control. ND: not 185 detected. 186 187 Farm ID +AIV Sample ID (10 copies/uL) Farm 1 + Sample 1 34.8 35.8 35.4 37.0 Farm 2 + Sample 1 30.7 30.7 34.7 35.3 Farm 3 + Sample 1 29.8 29.8 32.3 31.6 Farm 4 + Sample 1 31.3 31.2 33.2 33.0 Farm 5 + Sample 1 30.2 30.2 34.6 36.1 Farm 6 + Sample 2 30.7 31.3 33.0 32.5 Farm 7 + Sample 2 31.0 30.7 32.8 32.8 Farm 8 + Sample 2 30.9 31.1 31.8 32.2 Farm 9 + Sample 2 32.0 31.4 32.5 30.8 Farm 10 + Sample 2 31.3 31.4 32.4 32.5 Farm 11 + Sample 3 31.4 31.5 35.1 35.7 Farm 12 + Sample 3 31.2 31.4 33.4 33.0 Farm 13 + Sample 3 32.2 32.2 35.8 36.6 Farm 14 + Sample 3 31.6 31.7 31.0 31.3 Farm 15 + Sample 3 31.4 30.8 35.4 35.5 PC 24.9 25.1 24.5 23.7 NTC ND ND ND ND RNeasy + QS 5 (USDA benchtop) Ezextractor + MiniQ (MT) Ct values of two PCR assays .CC-BY-NC-ND 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted June 5, 2025. ; https://doi.org/10.1101/2025.06.02.657307doi: bioRxiv preprint 7 188 Figure 1. Comparison of performance of the MT vs. the benchtop PCR testing platforms. The 189 Ct values listed in Table 3 from the two platforms respectively are analyzed. The Spearman’s 190 correlation is calculated to be 0.96 and a paired Student’s t test is calculated to be 0.0002. The 191 difference of the mean Ct values of the MT assay and the USDA-benchtop assay is calculated 192 to be -2.3. 193 194 Sensitivity of the MT platform 195 Raw milk samples from Farm 8 and Farm 9, undergone serial dilutions, were 196 diagnosed using the MT platform to determine its detection sensitivity and efficiency. 197 The Ct values are listed in Table 4. The Ct values of two independent runs of each 198 sample at the five different dilution levels with estimated virus density of 1000, 100, 199 10, 1.0, and 0.10 copies/µL are consistent with each other and show the expected 200 trend according to the virus density. The Ct values as a function of dilution level, the 201 PCR standard curves in Figure 2, show that the amplification efficiencies are 93.6% 202 and 102.5% for Farms 8 and 9 respectively. The R 2 values (0.987 and 0.989) of both 203 dilution series show good PCR performance of the MT platform. 204 205 Farm ID/Dilution factor Ct values of the dilution series Farm 8/101 24.1 23.9 Farm 8/102 26.1 26.1 Farm 8/103 31.1 31.1 Farm 8/104 33.7 34.0 Farm 8/105 38.0 37.1 .CC-BY-NC-ND 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted June 5, 2025. ; https://doi.org/10.1101/2025.06.02.657307doi: bioRxiv preprint 8 Farm 9/101 23.7 23.6 Farm 9/102 27.4 27.3 Farm 9/103 31.3 31.1 Farm 9/104 34.0 34.2 Farm 9/105 36.1 37.0 PC 25.1 25.1 NTC ND ND Table 4 . Ct values of qPCR run using the MT platform of raw milk samples from two farms 206 spiked with inactivated AIV H5 following serial ten-fold dilutions. Two independent PCR 207 runs were performed for each sample. PC: positive control. NTC: no template control. ND: 208 not detected. Note: the first (10 1) dilution was made by adding 40 µL of inactivated virus 209 sample to 100 µL of milk and 260 µL PBS to achieve a virus density of 1000 copies /µL. 210 Subsequent ten-fold dilutions were made by adding 100 µL of prior dilution to 900 µL of 211 diluent described in Table 2. 212 Figure 2. PCR standard curves, Ct values as a function of dilution level, from the data in 213 Table 4. The amplification efficiency is calculated to be 93.6% for Farm 8 samples and 214 102.5% for Farm 9 samples. 215 216 Effect of Raw Milk Inhibitors on PCR Diagnosis 217 The data obtained using both platforms (Table 3) clearly show their effectiveness in 218 detecting inactivated AIV H5 in raw milk. Table 5 reports comparison of diagnoses 219 performed on raw milk vs. PBS samples spiked with the same virus density. All runs 220 on the internal positive control (Xeno™ RNA Control) spiked samples, whether they 221 contain raw milk or not, showed similar Ct values. In contrast, the runs on inactivated 222 viruses showed notable differences in Ct values: The runs on samples containing raw 223 milk showed consistently 1.5 higher Ct values than the pure PBS samples. Further, it 224 appears that whether the combination of Ezextractor+QS5 or Ezextractor + MiniQ, 225 the Ct values for both virus- and internal positive control- spiked samples appear the 226 same. 227 .CC-BY-NC-ND 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted June 5, 2025. ; https://doi.org/10.1101/2025.06.02.657307doi: bioRxiv preprint 9 228 Ezextractor + QS5 Ct values AIV H5 Internal positive control PBS 32.8, 32.6 30.7, 30.5 Raw milk + PBS 34.0, 34.0 30.0, 29.8 PC 28.1, 28.1 27.9, 27.9 NTC ND ND Ezextractor + MiniQ Ct values AIV H5 Internal positive control PBS 32.5, 32.5 29.5, 29.3 Raw milk + PBS 34.3, 34.0 30.1, 29.3 PC 27.8, 27.8 27.0, 26.8 NTC ND ND Table 5. Ct values of qPCR run on two raw milk samples in comparison with PBS solutions. 229 In both comparisons, the raw milk sample and the corresponding PBS solution were spiked 230 with the same unknown quantity of inactivated AIV H5. PC: positive control. NTC: no 231 template control. ND: not detected. 232 233 234

235 Milk is a complex mixture containing fat globules, proteins, sugar, and metal 236 ions. [13] Components in milk that may interfere with the AIV detection process 237 include calcium ions, collagen, and myoglobin. These ions/molecules can inhibit 238 DNA polymerase or reverse transcriptase, and Taq polymerase-degrading plasmin 239 [14]. To mitigate the inhibitory effect of raw milk, we first diluted milk samples in 240 PBS, and we were able to demonstrate the feasibility of detecting AIV H5 in diluted 241 raw milk samples using both platforms. Both qPCR systems (MiniQ and QS5) 242 exhibited consistent performance in detecting AIV H5 in raw milk samples. Virus 243 detection sensitivity remained robust despite the presence of PCR inhibitory 244 substances in raw milk. The results also illustrated the impact of milk from different 245 sources on AIV H5 detection. Table 3 shows that the Ct values from the MT platform 246 range from 30 to 32 for all samples except the raw milk from Farm 1 which has Ct 247 values higher than 35. The Ct values from the benchtop platform show a larger 248 variation from 31 to 37, again with the Farm 1 sample displaying the highest Ct 249 values. 250 251 Table 4 shows the sensitivity of qPCR detection of inactivated AIV H5 in raw 252 milk samples diluted to different virus densities. The five levels of dilution correspond 253 .CC-BY-NC-ND 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted June 5, 2025. ; https://doi.org/10.1101/2025.06.02.657307doi: bioRxiv preprint 10 to approximately 1,000, 100, 10, 1, 0.1 copies/µL. Even the lowest density AIV can 254 be detected by the MT platform with good PCR efficiency. The standard curves 255 quantitatively illustrate the precision of the MT platform diagnosis. 256 257 From the comparison of Ct values listed in Table 5, it seems that the inhibitors do 258 have effect on detection efficiency. The runs on samples containing raw milk show 259 consistently 1.5 higher Ct values than the pure PBS samples. Furthermore, by 260 comparing the experiments performed by using the same Ezextractor but different 261 qPCR (QS5 vs MiniQ), it appears that the difference in detection efficiency between 262 the bench top vs miniaturized instruments lies primarily with the extraction 263 method/instrument. 264 265 Examination of the Ct values in Table 3 also shows that the MT platform 266 displays higher sensitivity than the benchtop platform in this study. Ct values for all 267 15 raw milk samples from the MT platform are on average ~2 units lower than the 268 ones from the benchtop platform. This difference may be attributed to the different 269 extraction methods used in these two platforms. 270 271 272

273 In this study, we have shown that, even though the presence of the inhibitors in 274 raw milk challenges the operation of extraction and PCR detection, both the MT and 275 the benchtop platforms can detect inactivated AIV H5 in raw milk samples from 15 276 different farms. Further, the MT platform performs well and consistently for 277 inactivated virus samples diluted over a 5-order magnitude. The experiments 278 performed so far illustrate that AIV H5 can be detected with high sensitivity even 279 though there is discernible interference from milk components. The MT platform 280 performance in all samples tested in this work is more sensitive than the benchtop 281 platform used here. 282 283 The MT platform utilizing miniaturized extraction and PCR instruments for 284 inactivated AIV H5 detection in raw milk samples represents a highly sensitive, 285 versatile, and portable approach for managing AIV outbreaks. Successful containment 286 and mitigation of outbreaks at the earliest instance call for early detection and 287 surveillance of the virus. Given that viral loads in milk may vary depending on 288 infection stage and severity, the ability to detect the virus at low concentrations 289 ensures that even minimally infected herds can be identified and managed promptly. 290 This capability is crucial for preventing the spread of infection within and between 291 .CC-BY-NC-ND 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted June 5, 2025. ; https://doi.org/10.1101/2025.06.02.657307doi: bioRxiv preprint 11 herds. Putting into perspective the recent outbreak of H5N1 in the United States dairy 292 cattle and the associated zoonotic risks, the implementation of the MT platform offers 293 a practical solution for minimizing the public health and economic impacts. The 294 portability of these systems could facilitate rapid on-site testing, reducing the delay 295 between sample collection and results and communication with government officials. 296 This is particularly important in high-risk areas in many states which have 297 experienced significant outbreaks. 298 299 300 Data Availability. All data generated or analyzed in this study are included in this 301 article. PCR amplification curves of data shown in Table 3 and 4 are shown in Figure 302 3 and 4 respectively at the end of this article. 303 304

and Funding. We thank Dr. Tsun-Yung Kuo, Dr. Meei-Yun Lin, 305 Prof. Charles Chen, and Luke Tzu-Chi Liu of Schweitzer Biotech Co. and Dr. Barry 306 Arkles of DiaVac Biotech Co. for critical review and suggestions for the preparation 307 of the manuscript. DiaVac Biotech Co. provided the funding and Schweitzer Biotech 308 Co. provided the EZextractor Viral DNA/RNA Extraction Kit, EZextractor Nucleic 309 Acid Extraction System, and the EZcycler Mini real-time PCR (MiniQ) System for 310 this study. 311 312 Competing Interests. Authors declare no competing interests. 313 314 .CC-BY-NC-ND 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted June 5, 2025. ; https://doi.org/10.1101/2025.06.02.657307doi: bioRxiv preprint 12 315 316 317 318 Figure 3. Amplification plots of two qPCR runs for the detection of inactivated AIV H5 319 spiked into raw milk samples from 15 farms. The qPCR runs were performed using the MT 320 platform (EZextractor for RNA extraction and MiniQ thermocycler; right panels) or the 321 benchtop platform (RNeasy for RNA extraction and QS5 thermocycler; left panels). 322 323 324 325 326 327 328 329 .CC-BY-NC-ND 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted June 5, 2025. ; https://doi.org/10.1101/2025.06.02.657307doi: bioRxiv preprint 13 330 Figure 4. Amplification plots of two qPCR runs for the detection of AIV H5 in serial ten-fold 331 dilutions of inactivated AIV H5 spiked into raw milk samples from two farms. The qPCR runs 332 were performed using the MT platform (EZextractor for RNA extraction and MiniQ 333 thermocycler). 334 335 .CC-BY-NC-ND 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted June 5, 2025. ; https://doi.org/10.1101/2025.06.02.657307doi: bioRxiv preprint 14

336 1. Shi J, Zeng X, Cui P , Yan C, Chen H. Alarming situation of emerging H5 and 337 H7 avian influenza and effective control strategies. Emerg Microbes Infect 338 2023;12(1):2155072. doi: 10.1080/22221751.2022.2155072. 339 2. Graziosi G, Lupini C, Catelli E, Carnaccini S. Highly pathogenic avian 340 influenza (HPAI) H5 clade 2.3.4.4b virus infection in birds and mammals. 341 Animals 2024;14(9):1372. doi: 10.3390/ani14091372 342 3. Burrough ER, Magstadt DR, Petersen B, Timmermans SJ, Gauger PC, Zhang 343 J, Siepker C, Mainenti M, Li G, Thompson AC, Gorden PJ. Highly 344 pathogenic avian influenza A (H5N1) clade 2.3. 4.4 b virus infection in 345 domestic dairy cattle and cats, United States, 2024. Emerg Infect Dis 346 2024;30(7):1335. doi: 10.1128/mbio.02542-24. 347 4. Oguzie JU, Marushchak LV , Shittu I, Lednicky JA, Miller AL, Hao H, 348 Nelson MI, Gray GC. Avian influenza A (H5N1) virus among dairy cattle, 349 Texas, USA. Emerg Infect Dis 2024;30(7):1425. 350 doi:10.3201/eid3007.240717. 351 5. Uyeki TM, Milton S, Abdul Hamid C, Reinoso Webb C, Presley SM, Shetty 352 V , Rollo SN, Martinez DL, Rai S, Gonzales ER, Kniss KL. Highly 353 pathogenic avian influenza A (H5N1) virus infection in a dairy farm worker. 354 N Eng J Med 2024;390(21):2028-2029. doi: 10.1056/NEJMc2405371 355 6. Centers for Disease Control. Current Situation: Bird Flu in Dairy Cows. 356 Available at: Current Situation: Bird Flu in Dairy Cows | Bird Flu | CDC. 357 Accessed December 25, 2024. 358 7. Centers for Disease Control. H5 Bird Flu: Current Situation. Available at: H5 359 Bird Flu: Current Situation | Bird Flu | CDC. Accessed May 8, 2025 360 8. Dyer O. Bird flu: Canadian teenager is critically ill with new genotype. BMJ 361 2024;387:q2529. doi: 10.1136/bmj.q2529. 362 9. Centers for Disease Control. CDC Confirms First Severe Case of H5N1 Bird 363 Flu in the United States. Available at: 364 https://www.cdc.gov/media/releases/2024/m1218-h5n1-flu.html. Accessed 365 December 20, 2024. 366 10. Governor of California. Governor Newsom takes proactive action to 367 strengthen robust state response to Bird Flu. Available at: 368 https://www.gov.ca.gov/2024/12/18/governor-newsom-takes-proactive-action369 -to-strengthen-robust-state-response-to-bird-flu/. Accessed December 22, 370 2024. 371 11. U.S. Department of Agriculture. USDA Announces New Federal Order, 372 Begins National Milk Testing Strategy to Address H5N1 in Dairy Herds. 373 .CC-BY-NC-ND 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted June 5, 2025. ; https://doi.org/10.1101/2025.06.02.657307doi: bioRxiv preprint 15 Available at: 374 https://www.usda.gov/article/usda-announces-new-federal-order-begins-natio375 nal-milk-testing-strategy-address-h5n1-dairy-herds-0. Accessed December 22, 376 2024. 377 12. Animal and Plant Health Inspection Service, U.S. Department of Agriculture. 378 APHIS Requirements and Recommendations for Highly Pathogenic Avian 379 Influenza (HPAI) H5N1 Virus in Livestock For State Animal Health Officials, 380 Accredited Veterinarians and Producers. Available at: 381 https://www.aphis.usda.gov/sites/default/files/aphis-requirements-hpai-livest382 ock-eng-sp.pdf. Accessed December 23, 2024. 383 13. Haug A, Høstmark AT, Harstad OM. Bovine milk in human nutrition--a 384 review. Lipids Health Dis. 2007;6:25. doi: 10.1186/1476-511X-6-25 385 14. Schrader C, Schielke A, Ellerbroek L, Johne R. PCR inhibitors–occurrence, 386 properties and removal. J Appl Microbiol. 2012;113(5):1014-26. 387 .CC-BY-NC-ND 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted June 5, 2025. ; https://doi.org/10.1101/2025.06.02.657307doi: bioRxiv preprint