Ethics approval

All experiments were approved by the Animal Research Ethics Committee (CARE) of the Faculty of Veterinary Medicine, Vietnam National University of Agriculture (Approval No. CARE-2022/05). All procedures were performed in accordance with CARE guidelines and reported following the ARRIVE guidelines (https://arriveguidelines.org).

Epidemiological information and test results for individual pigs

Details are given in S-Table 1 of the supplementary materials.

Oral fluids sampling from raised pigs

As detailed in S-Table 1, a total of 68 pooled oral fluids samples from raised pigs were tested. Of these, 63 samples were collected at an early stage of disease from five farmers (Farms A–E) in four provinces in northern Vietnam between November 20, 2022 and August 25, 2023. On Farms A, D, and E, ASF occurred in one pen and subsequently spread to other pens; samples were collected on different days. On Farms B and C, pigs showed clinical symptoms suspicious for ASF one to two days after sampling. Additionally, five samples were obtained on November 12, 2023, from fattening pigs at five farmers in ASF-free Miyazaki Prefecture, Japan, and used as negative controls (Tables 1, 2 and S-Table 1). Oral fluids were collected using cotton ropes suspended in pens for 30 min²⁶, then transported to the lab. Samples were stored at − 80 °C (Vietnam) and − 30 °C (Japan) until use. Prior to testing, samples were thawed, centrifuged at 2000 g for 1–5 min and supernatants were transferred to 50 ml sterile plastic tubes. No more than 75% of the original volume was collected to avoid debris. For limit of detection (LOD) testing, ASFV-negative oral fluids (from four pig farms) were pooled to create at least 100 ml of matrix for spiking.

DNA extraction by reference method

To reduce viscosity and avoid clogging of extraction columns in accordance with the pathogen detection manual 2019-nCoV issued by the National Institute of Infectious Diseases, Japan (NIID-J, 2020)³⁴, 150 μl of oral fluids was diluted 1:3 with 450 μl of phosphate-buffered saline (PBS) in a 1.5 ml-sterile microcentrifuge tube. After vortexing, the diluted oral fluids were centrifuged at 15,600 g for 30 min, 150 μl of the supernatant was extracted using a DNeasy blood & tissue kit (Qiagen, Maryland, USA), and eluted in 50 μl. Alternatively, 200 μl of supernatant was extracted using the mag LEAD 6gC automated extraction platform (Precision System Science Co., Matsudo, Japan) and a reagent cartridge (MagDEA Dx SV, Precision System Science).

ASFV concentration by developed method

A previously described SARS-CoV-2 concentration method⁴ was adapted for ASFV. Briefly, 10 ml of oral fluid supernatant was transferred to a sterile 50 ml plastic centrifuge tube and diluted 1:3 with 20 ml of SAP (Semi-alkaline proteinase, Suputazyme; Kyokuto Pharmaceutical Industries, Tokyo, Japan). After vortexing, the tubes were incubated at room temperature for 15 min to digest oral fluid components. The mixture was then centrifuged at 15,557 g for 30 min. Subsequently, 24 ml of the supernatant was transferred to a new sterile 50 ml plastic tube and mixed with 9.6 ml of PEG-NaCl solution (equivalent to 40% of the total liquid volume³⁵). The tubes were vortexed and centrifuged again at 8000 g for 20 min to precipitate ASF virions. After the second centrifugation, the supernatant was carefully removed. Then, 100 µl (for use with the DNeasy Blood & Tissue Kit) or 150 µl (for use with the automated extraction platform) was added using a sterile pipette tip, and combined with the approximately 50 µl of residual liquid in the tube, bringing the final volume to approximately 150 μl or 200 μl.

To detach the ASF virion-PEG complex from the tube wall, the pellet (formed during the first centrifugation) was pipetted 10 times at the expected adhesion site using a 1 ml sterile long filtered tip to minimize contamination. A 1 ml sterile short tip was then used to vortex the sample, ensuring complete resuspension of the pellet. The resulting mixture containing the ASF virion–PEG complex was flushed into the tube, and DNA was extracted and purified in a final volume of 50 µl using either a column-based kit or an automated platform, as described above. For four samples (nos. 12, 21–23, shown in S-Table 1) with oral fluid volume was less than 10 ml, the volumes of SAP and PEG-NaCl solutions were adjusted proportionally to maintain the same mixing ratio.

Real-time PCR assay

Real-time PCR was performed on a CFX Opus 96 Real-Time PCR System (Bio-Rad Laboratories, Hercules, CA, U.S.A.) or a QuantStudio 3 (Thermo Fisher Scientific, Inc., Waltham, MA, U.S.A.) with two primers and probes consisting of the forward primer (5′-TGC TCA TGG TAT CAA TCT TAT CG-3′), the reverse primer (5′-CCA CTG GGT TGG TAT TCC TC-3′) and the probe (5′-FAM-TTC CAT CAA AGT TCT GCA GCT CTT-TAMRA-3′) reported by Tignon et al.³⁶. Amplification cycle conditions, reagent volumes and concentrations were slightly modified as follows. Briefly, 50 µl of real-time PCR reaction mixtures were used, comprising 25 µl of 2 × probe qPCR mix (Takara Bio Inc., Otsu, Japan), 0.2 µl each of forward and backward primers (100 µmol l⁻¹ each; Hokkaido System Science, Co., Ltd., Sapporo, Japan), 0.1 µl of probe (100 µmol l⁻¹; Hokkaido System Science), 19.5 µl of nuclease-free water (Takara Bio), and 5 µl of the DNA template. The cycling conditions were as follows: one cycle at 95 °C for 30 s, followed by 45 cycles each at 95 °C for 10 s, and 60 °C for 60 s. The automatically calculated Ct value was adopted, and the Ct cut-off value was set at 37.00.

Real-time LAMP assay

Real-time LAMP assay was performed following our previous report³⁰ with slight modifications. Briefly, in-house LAMP reaction mixtures were prepared by doubling the volume of reaction solution to 50 µl with the same composition as described. The amount of template DNA was set to 5 µl. The reaction mixtures were incubated at 63 °C for 45 min, followed by 80 °C for 5 min to complete the reaction using a real-time turbidimeter (Loopamp EXIA; Teramecs, Kyoto, Japan). When the derivative of turbidity reached 0.05 within 40 min, the reaction was automatically considered positive. Endpoint judgement by unaided eye was also performed. When the reagent color remained purple, the result determined as negative. On the other hand, a change to sky blue was interpreted as positive. For both real-time PCR and LAMP analyses, positive controls were DNA sequences derived from field isolates in Vietnam and artificial synthetic DNA sequences ordered from Eurofins Genomics K. K. (Tokyo, Japan) in Japan.

Preparation of tenfold dilution series of ASFV spiked oral fluids

The ASFV strain ASF/HY01/2019 (GenBank Accession no. MK554698; genotype II) was cultured on porcine alveolar macrophages (TCID₅₀ = 10⁶^.5), stored at − 80 °C, and subsequently diluted tenfold in PBS. The diluted virus stock was then stored at 4 °C. In parallel, pooled pig oral fluid samples stored at − 80 °C were thawed at room temperature. The supernatant, created by centrifugation at 2000 g for 5 min to remove cotton rope and feed residue, was transferred to two new 50 m-sterile tubes for the spike test. Next, the pig oral fluid supernatant was dispensed into six 50 ml tubes of 10 ml each, and then, the tenfold dilution series of ASFV was sequentially spiked and vortexed thoroughly. A series of pig oral fluids containing from neat to10⁻⁵-foldASFV dilutions was prepared, as shown in Table 4.

Determination of LODs

DNA extracted from each dilution using both methods was tested in duplicate by real-time PCR and LAMP. Results were interpreted as positive if both replicates were positive; otherwise, negative (Table 4).

Statistical analysis

Diagnostic sensitivity, diagnostic specificity, positive predictive value (PPV), and negative predictive value (NPV) were calculated (Tables 1, 2). Differences between paired proportions were evaluated using McNemar’s test in R (version 4.4.0³⁷). Statistical significance was set at p < 0.05 (Figs. 1, 2).