{"paper_id":"0dd7f53d-0c6b-4ae5-bfb6-a9003c177183","body_text":"1 \n \nIsothermal Detection of Influenza D using RT-LAMP \n \nCarlos Abelardo dos Santos1, Jialu Li1, Pauline M. van Diemen2, Andrew McMahon3, Benjamin \nC. Mollett2, Andrew M. Ramsay2, Meshach M. Maina 4, Joe James 2, Helen E. Everett 2, Janet \nM. Daly4 and Nicole C. Robb1* \n \n1Warwick Medical School, University of Warwick, CV4 7AL, United Kingdom \n \n2Animal and Plant Health Agency, APHA-Weybridge, New Haw, Addlestone, KT15 3NB, \nUnited Kingdom \n \n3School of Life Sciences, University of Warwick, CV4 7AL, United Kingdom \n \n4One Virology, Wolfson Centre for Global Virus Research, Sutton Bonington Campus, \nUniversity of Nottingham, LE12 5RD, United Kingdom \n \n*Nicole.Robb@warwick.ac.uk  \n.CC-BY 4.0 International licenseperpetuity. It is made available under a \npreprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in \nThe copyright holder for thisthis version posted January 9, 2026. ; https://doi.org/10.64898/2026.01.09.698603doi: bioRxiv preprint \n\n \n2 \n \nAbstract \nThe Orthomyxoviridae family includes influenza D virus (IDV), an emerging pathogen primarily \naffecting cattle and swine, with evidence of cross-species transmission and potential zoonotic \nrisk. Although active human infections have yet to been confirmed, high seroprevalence in \ncattle-exposed populations highlights the need for continued surveillance. Here, a rapid, field-\ndeployable RT-LAMP assay for IDV detection was developed and validated, with 99.2% \nspecificity and sensitivity ranging from 95.6% (Cq < 30) to 81.8% (Cq < 40). This method offers \na cost-effective, accessible alternative to RT-qPCR, enabling improved monitoring of IDV, and \nreinforcing preparedness for emerging influenza threats. \n \nIntroduction \nThe family Orthomyxoviridae comprises nine genera, four of which are influenza \nviruses. Influenza A (IAV) and influenza B (IBV) are the main viruses responsible for human \nseasonal influenza epidemics, whilst IAV poses a pandemic threat linked to bidirectional \ntransmission between animals to humans. Both IAV and IBV can cause severe illness in \nhumans, whilst influenza C virus (ICV) infection is associated with milder upper respiratory \ntract symptoms, most commonly in children under the age of 2. Influenza D virus (IDV) \nwas first discovered in 2011 in pigs in Oklahoma, USA, and subsequently found in cattle, which \nare now considered the main livestock reservoir. In cattle, IDV infection is associated with \nbovine respiratory disease (BRD) as the primary viral infection, which may predispose cattle \nto opportunistic secondary bacterial infections (1). \n Whilst the majority of IDV occurrences have been in swine and cattle, IDV antibodies \nhave also been found in other domesticated animals such as sheep (2), goats (3), horses (3), \nwild boars (4) and camelids (5), suggesting that cross-species transmission may occur. Mice, \nferrets and guinea pigs have also been found to be susceptible to experimental viral infection \n(6). Similarly to ICV, IDV has been shown to bind to sialic acid receptors on the cell surface of \nthe host, specifically 9-O-acetylated sialic acid receptors (7). These receptors are found \nthroughout the entire respiratory tract in cattle, as well as in the nasal and pharyngeal \nepithelium of pigs, sheep, goats, and horses, suggestive of a potentially wide host range (8). \nInfluenza D virus has been shown to propagate effectively in various human cell types \n(9), however, no active infection of humans has been reported to date, and although IDV was \ndetected in a nasal wash sample from a swine farm worker in Malaysia, no infectious virions \nwere retrieved (10). The presence of IDV antibodies has been the primary method of inferring \npast exposure to IDV in humans. A 2011 study found a 1.3% seroprevalence of IDV antibodies \nin the general human population in the USA and Canada (11). Another study in Italy discovered \nthat the seroprevalence of IDV in the general human population increased from 5% in 2005 to \n46% in 2014 (12). In human cohorts occupationally exposed to cattle, the prevalence of IDV \n.CC-BY 4.0 International licenseperpetuity. It is made available under a \npreprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in \nThe copyright holder for thisthis version posted January 9, 2026. ; https://doi.org/10.64898/2026.01.09.698603doi: bioRxiv preprint \n\n \n3 \n \nantibodies reached as high as 97%(13), reflecting a similar prevalence of antibodies in cattle \n(14). \nAlthough IDV infection has not been observed to induce severe disease in humans, it \nremains crucial to monitor its prevalence due to the ability of influenza viruses to evolve and \ntransmit across species barriers. While antibody detection serves as a robust epidemiological \ntool on a population scale (15), reverse transcriptase quantitative polymerase chain reaction \n(RT-qPCR) stands as the gold standard for detecting viral RNA and thus active infection at an \nindividual level. However, the necessity for specialized laboratories and trained personnel to \nprocess and analyse the samples, coupled with high costs and limited utility for field-based \ntesting, poses a significant challenge to high throughput. Reverse transcription loop-mediated \nisothermal amplification (RT-LAMP) is a rapid, highly sensitive, and cost-effective molecular \ndiagnostic tool that has emerged as a potential alternative to PCR-based methods (16). RT-\nLAMP can be performed at a constant temperature and return a result in less than an hour, \nmaking it well suited for point-of-sample testing. The success of RT-LAMP during the COVID-\n19 pandemic underscores its potential as a valuable tool in our ongoing efforts to combat \ninfectious diseases (17). \n Here, we report an RT-LAMP assay to specifically detect IDV with particular application \nfor field setting diagnostic assays in resource-poor settings. We tested a total of 28 spiked \nsamples and 46 isolated RNA samples from swine and cattle using the assay, which achieved \na specificity of 99.23% and a sensitivity ranging from 95.60% for samples with a Cycle \nquantification (Cq) < 30, to 81.82% for all samples up to a Cq of 40. This development not only \nenhances our ability to detect and monitor IDV but also highlights the potential of RT-LAMP \nas a valuable tool in monitoring infectious diseases. We believe that our findings will support \nongoing efforts to mitigate the risks associated with this virus and enhance our collective \ncapacity to respond to infectious diseases. \n \nMaterials & Methods \n \nVirus Isolates \nThe isolates D/swine/Oklahoma/1334/2011 (11) and D/bovine/France/5920/2014 \n(18,19) were propagated in embryonated chicken eggs and on hRT18G cells respectively. \nBoth strains are well characterized and produced with a minimum number of passages to \navoid loss of strain fidelity. D/swine/England/126471/2023 is a more recent isolate and was \npropagated on ST cells (20). \n \nRNA extraction \n.CC-BY 4.0 International licenseperpetuity. It is made available under a \npreprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in \nThe copyright holder for thisthis version posted January 9, 2026. ; https://doi.org/10.64898/2026.01.09.698603doi: bioRxiv preprint \n\n \n4 \n \n We carried out the pre-isolation steps depending on the nature of the sample. For viral \nisolates, 20 L of the cell supernatant was lysed with 350 L of Monarch StabiLyse DNA/RNA \nBuffer, and we carried out the extraction according to the manufacturer’s protocol. We also \nextracted RNA from a swine lung fragment by excising a 10 mg tissue fragment and lysing it \nin 200 L of StabiLyse DNA/RNA Buffer and 200 L of nuclease-free water (Integrated DNA \nTechnologies) in a FastPrep-24 bead beater homogenizer (MP Biomedicals) for approximately \n10 minutes. We extracted total RNA from each preparation using the Monarch Spin RNA \nIsolation Kit (New England Biolabs), according to the manufacturer’s instructions. \n \nRT-qPCR assay for IDV \nThe RT-qPCR primers and probes for IDV detection used in this study were originally \npublished by Faccini et al. (21). The reactions were performed in a 384-well Quant Studio 5 \nReal-Time PCR System (Applied Biosystems) using the Luna Universal Probe One-Step RT-\nqPCR Kit (New England Biolabs) following the manufacturer’s instructions. The primers and \nprobe target the PB1 gene (from position 1215 to 1323, Accession number JQ922306), so we \ndesigned a PB1 gBlock to use as control containing the nucleotides from position 500 to 1500. \nThe gBlock was synthesized by Integrated DNA Technologies. We eluted the gBlock according \nto the manufacturer’s protocol and confirmed the concentration by UV absorbance using a \nNanoDrop spectrophotometer (Thermofisher Scientific). We used serial dilutions ranging from \n10^6 to 10 copies of the IDV gBlock to generate a standard curve, which was employed to \nquantify the RNA material from isolates, mock samples and samples. \n \nRT-LAMP Primer Design \nWe downloaded all the PB1 and PB2 complete gene sequences of IDV available from \nthe National Centre for Biotechnology Information (NCBI) database (accessed December \n2024) and generated alignments using the MAFFT multiple sequence alignment tool \n(https://mafft.cbrc.jp/alignment/software/). We designed the primers for RT-LAMP against PB1 \nand PB2 using the PrimerExplorer V5 software \n(https://primerexplorer.eiken.co.jp/lampv5e/index.html). To improve coverage and guarantee \namplification of the maximum number of sequences, we set the consensus threshold to 90%. \nWe marked all the mutated positions in the primer design software and created the primers \nusing the “Common” design option, which allows for primers to be designed targeting mutated \nregions if the position of the variable nucleotide within the primer is unlikely to affect \namplification efficiency. Primers were filtered by end stability, and the maximum G for the 3’ \nend of F3/B3, F2/B2, LF/LB and the maximum G for the 5’ end of F1c/B1c was set to be ≤ -\n4.00 kcal/mol. Primers were also checked for the presence of stable hairpins, self-dimers and \n.CC-BY 4.0 International licenseperpetuity. It is made available under a \npreprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in \nThe copyright holder for thisthis version posted January 9, 2026. ; https://doi.org/10.64898/2026.01.09.698603doi: bioRxiv preprint \n\n \n5 \n \nhomodimers using IDT’s Oligo Analyzer tool \n(https://eu.idtdna.com/pages/tools/oligoanalyzer). Sets containing primers with stable hairpins \nor dimers were not synthesised. The primers FIP and BIP are composed of two different \nregions (F1c/F2 for FIP and B1c/B2 for BIP), which mediate the creation of the loops \ncharacteristic of LAMP . Target recognition by LAMP primers creates an ever-growing \nconcatemer containing single-stranded loops that also serves as target for continuous \namplification. To decrease local Tm in the loops and lower the chances of undesired secondary \nstructures that could hinder amplification, we inserted T linkers (22) between the F1c/F2 and \nB1c/B2 regions in the FIP/BIP primers. \n \nColorimetric one-step RT-LAMP assay for IDV \n We prepared all reactions as a total volume of 20 L, each containing a final \nconcentration of 1X WarmStart® Colorimetric LAMP Master Mix with UDG (Uracil DNA \nGlycosylate) (New England Biolabs), 1X primer mix (containing 0.2 mM F3/B3, 1.6 mM \nFIP/BIP and 0.8 mM LF/LB primers), 10% (v/v) of target, and nuclease-free water (Integrated \nDNA Technologies). For real-time analysis, we added 1 mM of SYTO9 (Thermofisher \nScientific) to the reaction mixture. To optimize the assay, we incubated the reactions in a real-\ntime thermocycler (QuantStudio 3, Applied Biosystems) at 65°C for 80 cycles, each cycle \nconsisting of 45 seconds of incubation followed by fluorescence read in the FAM channel for \n15 seconds. Subsequent reactions were incubated in a conventional T-100 PCR instrument \n(Bio-Rad) at 65°C for 60 minutes. Unless otherwise specified, all reactions were performed in \ntriplicate.  \n \nSensitivity and specificity of the Influenza D assay \nTo assess whether we could differentiate IDV from closely-related viruses, we tested \nthe RT-LAMP assay for IDV against gBlocks containing the PB2 target gene segment of all \nother influenza strains. While the gBlock containing the complete PB2 gene for IDV \n(D/swine/Oklahoma/1334/2011) was used as positive control in our reactions, gBlocks \ncontaining the complete PB2 gene for Influenza A (IAV, A/Puerto Rico/8/1934), Influenza B \n(IBV, B/Lee/1940), and Influenza C (ICV, C/Ann Arbor/1/50) viruses were used as negative \ncontrols. We carried out the reactions using either 2 L of the respective gBlocks controls or \nthe equivalent amount of nuclease-free water as a non-template control (NTC). All the gBlocks \nwere tested at a concentration of 10 6 copies per L for 1 hour. Results were analysed in \ntriplicate, based on the colorimetric change in the reaction from pink (negative) to yellow \n(positive). \n.CC-BY 4.0 International licenseperpetuity. It is made available under a \npreprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in \nThe copyright holder for thisthis version posted January 9, 2026. ; https://doi.org/10.64898/2026.01.09.698603doi: bioRxiv preprint \n\n \n6 \n \n To determine the sensitivity of the assay, we mixed equimolar amounts of extracted \nRNA from each of the three viral isolates (D/bovine/France/5920/2014, \nD/swine/Oklahoma/1334/2011 and D/swine/England/126471/2023), which had been \nquantified by RT-qPCR using the assay described previously. We serially-diluted the RNA mix \nand verified the concentrations by RT-qPCR. We tested the serially diluted RNA mix by RT-\nLAMP in octuplicate and fitted a Probit regression curve (dose-response plot) using the IBM \nSPSS Statistics Software (Version 29.0.2.0). We defined the limit of detection to be equal to \nthe minimum number of copies detected in at least 95% of the reactions. \n We analysed the amplification products on a 2% agarose gel in 1X TBE (Tris-Borate-\nEDTA) Buffer using SYBR Safe DNA Gel Stain (Thermofisher) and the Quick-Load 100 bp \nDNA Ladder (New England Biolabs). The gel was run at 100V for 60 minutes and images were \nacquired with a ChemiDoc Imaging System (BioRad). \n \nEvaluation using spiked and field-collected animal samples \nTo assess the performance of the test in a more complex  sample matrix, we prepared \n28 spiked samples by mixing total purified RNA extracted from swine lung tissue with variable \namounts of viral RNA from the three virus isolates used in this study. Samples were quantified \nby RT-qPCR and tested by RT-LAMP . Results were analysed by the colour change from pink \n(negative) to yellow (positive) in the reaction after 60 minutes of incubation.  \nTo assess the viability of the assay in real animal samples, we analysed 46 swine and \ncattle RNA samples. These samples comprised nasal swab and/or respiratory tissue samples \n(lung, viscera) from pigs and cattle with respiratory disease signs (cough and/or dyspnoea) \nsubmitted through the passive APHA surveillance networks, and containing a mixture of IAV \nor IDV positive or negative samples   \nThe specificity and sensitivity of the RT-LAMP assay for detecting IDV was assessed \nby comparing its results with those of the RT-qPCR assay, using the free online tool Medcalc \n(available at https://www.medcalc.org/calc/diagnostic_test.php).  \n \nResults \nOptimization of an RT-LAMP assay for Influenza D  \nThe primer sets generated for the PB1 and PB2 gene sequences of IDV were \nevaluated in silico for their end stabilities, formation of hairpins, self-dimers and heterodimers. \nWe found that none of the sets targeting the PB1 gene passed our filters, while two primer \nsets targeting the PB2 gene met our criteria and were synthetized. The sequences of the two \nRT-LAMP primer sets that passed our filters are presented in Table S1.  \nWe initially trialled both primer sets by performing the RT-LAMP reaction in a real-time \nthermocycler targeting gBlocks containing the complete PB2 sequence of IDV, as well as IAV, \n.CC-BY 4.0 International licenseperpetuity. It is made available under a \npreprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in \nThe copyright holder for thisthis version posted January 9, 2026. ; https://doi.org/10.64898/2026.01.09.698603doi: bioRxiv preprint \n\n \n7 \n \nIBV and ICV as negative controls, using SYTO9 DNA dye to track the accumulation of \namplification products over time (Figure 1). Each gBlock was incubated at a concentration of \n106 copies per µL of input. As non-specific amplification is a well-known problem with RT-\nLAMP reactions, we incubated the reactions for up to 80 minutes to assess the maximum \nincubation time for each primer set, defined as the longest time in which reactions could \nproceed without producing false-positive signals in the negative samples or non-template \ncontrols. While both primer sets amplified the Influenza D target gBlock at around the 15 \nminute mark, the IDV_set1 showed early onset of non-specific amplification at around 40 \nminutes (Figure 1A), compared to 65 minutes with the IDV_set2 (Figure 1B). Due to the \nincreased specificity, IDV_set2 was therefore chosen for all subsequent analysis. \nTo ensure complete coverage of at least 90% of the published IDV lineages, including \nthe three viral isolates that we had access to (D/bovine/France/5920/2014, \nD/swine/Oklahoma/1334/2011 and D/swine/England/126471/2023), degenerate nucleotides \nwere incorporated into the primers from IDV_set2 where necessary (Table S1 and Figure 2A). \nTo test whether the primer set containing degenerate nucleotides enabled detection of all three \navailable isolates of IDV, we incubated dilutions of extracted RNA at 104 and 103 copies/µL for \neach of the isolates. All concentrations of the three isolates were efficiently amplified after \napproximately 15 minutes (Figure 2B).  \n \nThe colorimetric RT-LAMP assay exhibits high sensitivity and specificity  \nHaving shown that the RT-LAMP primers efficiently detected IDV RNA using a real-\ntime thermocycler, we incubated subsequent reactions in a conventional thermocycler and \nanalysed the results visually based on a colorimetric change in the reaction from pink \n(negative) to yellow (positive).  \nTo show that the addition of the degenerate nucleotides did not impact primer \nspecificity, a high number of gBlock copies (106 copies/µL) of the PB2 gene of influenza A, B, \nC and D were incubated with the RT-LAMP mixture for 60 minutes at 65°C, only the IDV target \nwas amplified (Figure 3A), suggesting that our assay is specific to IDV. This result was \nconfirmed via gel electrophoresis of the amplicons (Figure 3B). \nNext, the sensitivity of the essay was estimated from the amplification results of eight \nreplicates at low target concentrations (500, 250, 100, 50 and 10 copies per µL of input) using \nan equimolar mix containing the RNA from the three IDV isolates used in this study. The \ncalculated probit regression curve was plotted and the minimum number of copies detected \n95% of the time was estimated to be 167 copies/µL (95% C.I. 103–3115 copies/µL) (Figure \n3C&D).  \n \nEvaluation of RT-LAMP assay on spiked and  diagnostic samples \n.CC-BY 4.0 International licenseperpetuity. It is made available under a \npreprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in \nThe copyright holder for thisthis version posted January 9, 2026. ; https://doi.org/10.64898/2026.01.09.698603doi: bioRxiv preprint \n\n \n8 \n \nTo assess the efficiency of the RT-LAMP test in a more complex matrix, we spiked \nvarying concentrations of IDV RNA into total RNA extracted from swine lung tissue. Prior to \naddition of the IDV RNA the swine lung tissue was confirmed to be free of IDV RNA via RT-\nqPCR. Twenty-seven spiked samples were tested in duplicate in different amounts of swine \nlung RNA to mimic different target/background ratios, using 4 non-spiked samples as negative \ncontrol (a total of 58 individual RT-LAMP reactions). The IDV RNA in all spiked samples was \nalso quantified by RT-qPCR. The RT-LAMP assay successfully detected all samples that had \na Cycle quantification (Cq) number lower than 29. It failed to amplify 2 samples out of 16 with \na Cq between 29 and 30, 4 samples out of 14 with a Cq between 30 and 32 and 3 samples \nout of 4 with a Cq over 32 (Figure 4A). All uninfected swine lung RNA samples were negative. \nFinally, we tested the performance of the RT-LAMP assay in detecting IDV viral RNA \nin total RNA extracted from samples collected from animals in the field. We analysed a total \nof 46 RNA samples extracted from cattle and swine. The samples were composed of negative \nsamples, samples containing IDV and samples containing IAV RNA. They were tested in \ntriplicate and the results compared to RT-qPCR as shown in Table 1. The RT-LAMP assay \nsuccessfully detected IDV RNA in 3 out of the 4 animal samples that were positive for IDV \nRNA by RT-qPCR. \nAltogether, we tested a total of 196 individual RT-LAMP reactions, comprising 58 \nspiked RNA and 138 animal RNA samples. We compared the RT-LAMP results to the results \nof the gold standard RT-qPCR assay, and calculated the sensitivity, specificity, positive and \nnegative likelihood ratio, and positive and negative predictive values. Given that amplification \nof low abundance targets close to the limit of detection are probabilistic events, multiple \nreplicates are required to estimate the detection probability (or hit rate) at a given \nconcentration. To account for this, every replicate of a spiked and diagnostic sample was \ntreated as an independent observation for statistical analysis, and each result was individually \ncompared to RT-qPCR. A total of 54 of the reactions were positive both by RT-qPCR and RT-\nLAMP (true positive), 129 were negative by both RT-qPCR and RT-LAMP (true negative), 12 \nwere positive by RT-qPCR and negative by RT-LAMP (false negative) and 1 was considered \npositive by RT-LAMP while it was negative by RT-qPCR (false positive) (Table S2). The \ncalculated specificity of the test was 99.23%, the sensitivity of the test was 81.82% and the \noverall accuracy was 93.37%. The sensitivity of the test for samples with a Cq < 35 was 85.7% \nand with a Cq < 30 was 95.6% (Figure 4B and Tables S3).  \n \nDiscussion \n We have developed a colorimetric RT-LAMP assay capable of detecting IDV RNA in \nboth experimentally spiked and field-collected samples. The assay demonstrated a high \nanalytical specificity of 99.23% as well as analytical sensitivity of 167 viral RNA copies per µL \n.CC-BY 4.0 International licenseperpetuity. It is made available under a \npreprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in \nThe copyright holder for thisthis version posted January 9, 2026. ; https://doi.org/10.64898/2026.01.09.698603doi: bioRxiv preprint \n\n \n9 \n \nof target input within a 60-minute reaction time, utilizing a one-step assay and a simple \ncolorimetric detection system for result interpretation. The single tube format allows for simpler \nhandling as fewer manipulations are required, decreasing time from collection to final result. \nActive viral infections typically produce millions of viral copies, therefore the detection limit in \nthe hundreds of copies range (equivalent to Cq values >30 in RT-qPCR assays) will enable \nidentification of infections during early-stage viral replication and throughout extended viral \nshedding periods. While sensitivity remains lower than RT-qPCR methodologies, the assay \nexceeds the analytical performance of most rapid antigen tests and could provide direct \nevidence of active infection compared to serological approaches.  \nPrimer design incorporated strategically positioned degenerate nucleotides to ensure \ncomprehensive viral lineage coverage, achieving detection capability for at least 90% of \npublished IDV sequences, including the recently identified D/swine/England/126471/2023 \nlineage from UK swine populations (20). The degenerate nucleotides were specifically \nintegrated avoiding the ends of the primers to minimize potential amplification interference \nwhile maximizing lineage inclusivity. \nCurrent IDV detection relies mainly on RT-qPCR and antibody detection, both \npresenting distinct analytical trade-offs that impact surveillance implementation strategies. RT-\nqPCR is the gold-standard method and offers superior analytical sensitivity with viral load \nquantification, enabling precise molecular characterization essential for genetic surveillance \nand outbreak investigation. However, this approach requires sophisticated laboratory \ninfrastructure, specialized personnel, and complex sample preparation protocols. This limits \npoint-of-care applications and increases costs per sample. Conversely, antibody sampling \nprovides simplified collection protocols with enhanced sample stability and cost-effective \npopulation seroprevalence assessment, facilitating large-scale epidemiological studies. \nNevertheless, serological methods exhibit inherent limitations due to seroconversion kinetics, \ncannot distinguish active from historical infections, and may demonstrate cross-reactivity with \nrelated viruses, precluding their utility for acute infection diagnosis and real-time outbreak \nresponse. These complementary yet functionally distinct methodological constraints \nnecessitate the development of alternative diagnostic platforms that bridge the gap between \nlaboratory-based precision and field-deployable practicality. \nRT-LAMP presents a promising point-of-care alternative to RT-qPCR, as its simplified \nequipment requirements and minimal reaction components enable field-deployable testing \nwith reduced sample processing demands. The colorimetric RT-LAMP approach provides \nqualitative detection results, facilitating rapid screening of larger sample volumes at reduced \nper-reaction costs compared to conventional RT-qPCR methodologies. Even though RT-\nLAMP may be used to differentiate between known lineages, RT-LAMP cannot fully substitute \nRT-qPCR or sequencing for comprehensive genetic surveillance applications, as its \n.CC-BY 4.0 International licenseperpetuity. It is made available under a \npreprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in \nThe copyright holder for thisthis version posted January 9, 2026. ; https://doi.org/10.64898/2026.01.09.698603doi: bioRxiv preprint \n\n \n10 \n \namplification products are difficult to sequence. This is particularly problematic when detecting \nnovel or uncharacterized viral lineages that have not been sequenced and deposited in public \ndatabases. Despite these genetic surveillance limitations, RT-LAMP offers significant potential \nfor cost-effective sample triage protocols, enabling preliminary screening before confirmatory \nmolecular characterization through more sophisticated diagnostic platforms. \nIn summary, our RT-LAMP assay for IDV demonstrated high specificity and sensitivity, \nindicating its utility as a rapid and accessible diagnostic tool. To further enhance its \napplicability, future optimisations could focus on integrating the assay into lab-on-a-chip \nmicrofluidic platforms or coupling the reaction with a compact detector, allowing for fully \nautomated, portable testing. Further field validation across varied sample types and conditions \nwill be essential, however embedding this RT-LAMP assay into existing surveillance \ninfrastructures can bolster real-time IDV detection and strengthen One Health strategies for \npandemic preparedness. \n \nAcknowledgements \nThe Isolate D/bovine/France/5920/2014 was kindly provided by Dr Mariette Ducatez, Ecole \nVétérinaire de Toulouse (ENVT). This work was supported by a Royal Society Dorothy \nHodgkin Research Fellowship [DKR00620 to N.R.] and an Institute for Global Pandemic \nPlanning (IGPP) funded PhD at the University of Warwick, UK [to C.A.dS]. Influenza research \nat APHA is supported by DEFRA and the devolved Scottish and Welsh Governments \npreviously through FluFutures2 (SE2213) and currently through FluFocus (SE2227). \nSurveillance sample submissions to APHA were obtained under the National DEFRA-funded \nsurveillance programs SV3041 (swine influenza) and ED1000 and ED200 (bovine respiratory \nvirus). \n  \n.CC-BY 4.0 International licenseperpetuity. It is made available under a \npreprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in \nThe copyright holder for thisthis version posted January 9, 2026. ; https://doi.org/10.64898/2026.01.09.698603doi: bioRxiv preprint \n\n \n11 \n \n \n \n \nFigure 1.  Mean amplification curves from RT-LAMP reaction triplicates showing the \nonset of specific and non-specific signal over time . RT-LAMP on gBlocks containing the \ninfluenza A, B, C and D virus PB2 gene sequences. All gBlocks were tested at a concentration \nof 106 copies per µL of input. A) Reactions using the primers in IDV_set1. The amplification of \nthe IDV gBlock positive control started within the first 17 minutes, whilst IAV, IBV and ICV \ngBlocks and the NTC produced false-positive signals after only 35 minutes of incubation.  B) \nReactions using the primers in IDV_set2. The amplification of the IDV gBlock positive control \nstarted within the first 15 minutes, whilst IAV, IBV and ICV gBlocks and the NTC produced \nfalse-positive signals after 65 minutes of incubation. NTC: Non-template control (water). \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n.CC-BY 4.0 International licenseperpetuity. It is made available under a \npreprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in \nThe copyright holder for thisthis version posted January 9, 2026. ; https://doi.org/10.64898/2026.01.09.698603doi: bioRxiv preprint \n\n \n12 \n \n \n \nFigure 2. RT-LAMP primers targeting the influenza D virus (IDV) PB2 gene efficiently \namplify viral isolate RNA. A) All available IDV PB2 sequences in the NCBI database were \naligned and mutations were annotated. To improve coverage, primers were synthesised with \ndegenerate nucleotides in positions with heterogeneity in the sequences. The consensus \nsequence and sequences of the three isolates, D/bovine/France/5920/2014, \nD/swine/Oklahoma/1334/2011, and D/swine/England/126471/2023 are shown for \ncomparison. Primer target sequences are presented within the boxes, and the direction of \nbinding is represented by the arrows. B) Mean amplification curves from RT-LAMP reaction \ntriplicates showing the onset of specific and non-specific signal over time. All IDV isolate RNAs \nat 104 and 10 3 copies per  µL amplified within the first 20 minutes regardless of the lineage. \nNTC: Non-template control (water).  \n \n.CC-BY 4.0 International licenseperpetuity. It is made available under a \npreprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in \nThe copyright holder for thisthis version posted January 9, 2026. ; https://doi.org/10.64898/2026.01.09.698603doi: bioRxiv preprint \n\n \n13 \n \n \nFigure 3. Specificity and sensitivity assays for influenza D virus (IDV) detection using \nRT-LAMP . A) The total of 106 copies/µL of gBlocks containing the influenza A, B, C and D virus \nPB2 gene sequences were tested. Only the gBlocks containing the target sequences were \namplified as seen by the colour change in the reactions. NTC: non-template control (water). \nB) 2% gel electrophoresis of one representative of each amplification products of the RT-\nLAMP reactions shown in A), showing the typical lamp ladder-like pattern in IDV virus samples \nand no visible non-specific bands in the other target samples. C) Dilutions of target RNA in \neight replicates for sensitivity analysis at different concentrations, varying from 500 copies per \nµL (cp/l) of target (top row) to 10 copies per µL of target (bottom row).  D) The probit-analysis \ncurve of probability of detection in relation to the number of copies. The limit of detection is \ndefined as the minimum number of copies detected >95% of the time. \n \n.CC-BY 4.0 International licenseperpetuity. It is made available under a \npreprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in \nThe copyright holder for thisthis version posted January 9, 2026. ; https://doi.org/10.64898/2026.01.09.698603doi: bioRxiv preprint \n\n \n14 \n \n \n \nFigure 4. Sensitivity of the test in relation to RT-qPCR cycle quantification. A) Detection \nof spiked samples detected by RT-LAMP in relation to the cycle quantification (Cq) value. B) \nEach true positive sample was sorted according to its Cq value and RT-LAMP result. Yellow \nlines presented along the top line are positive by both RT-LAMP and RT-qPCR, pink lines on \nthe lower line are positive by RT-qPCR but negative by RT-LAMP. The sensitivity of each range \nis presented as percentages at the top of the graph. \n \n \n \n \n \n \n \n \n \n \n \n \n.CC-BY 4.0 International licenseperpetuity. It is made available under a \npreprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in \nThe copyright holder for thisthis version posted January 9, 2026. ; https://doi.org/10.64898/2026.01.09.698603doi: bioRxiv preprint \n\n \n15 \n \nTable 1. Comparison between RT-qPCR and RT-LAMP results for animal RNA samples. \nIDV: Influenza D Virus, IAV: Influenza A Virus. Samples 1-32 are from swine and 33-46 are \nfrom cattle. * Sample number 40 showed a change in colour in one of the replicates that did \nnot fully convert into yellow, but for statistical calculations it was considered a false positive. \n \nSample \nIDV  \nRT-LAMP \nIDV \nRT-qPCR \nIAV \nRT-qPCR \n1 Negative Negative Negative \n2 Negative Negative Positive – Cq 36.45 \n3 Negative Negative Negative \n4 Negative Negative Negative \n5 Negative Negative Negative \n6 Negative Negative Negative \n7 Negative Negative Negative \n8 Negative Negative Positive – Cq 39.24 \n9 Negative Negative Negative \n10 Negative Negative Negative \n11 Negative Negative Negative \n12 Negative Negative Negative \n13 Negative Negative Negative \n14 Negative Negative Positive – Cq 32.19 \n15 Negative Negative Negative \n16 Positive Positive – Cq 22.29 Negative \n17 Negative Negative Negative \n18 Negative Negative Negative \n19 Negative Negative Negative \n20 Negative Negative Positive – Cq 27.88 \n21 Negative Negative Negative \n22 Negative Negative Negative \n23 Negative Negative Negative \n24 Negative Positive – Cq 37.09 Negative \n25 Negative Negative Negative \n26 Negative Negative Negative \n27 Negative Negative Negative \n28 Negative Negative Negative \n29 Negative Negative Negative \n.CC-BY 4.0 International licenseperpetuity. It is made available under a \npreprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in \nThe copyright holder for thisthis version posted January 9, 2026. ; https://doi.org/10.64898/2026.01.09.698603doi: bioRxiv preprint \n\n \n16 \n \n30 Negative Negative Negative \n31 Negative Negative Negative \n32 Negative Negative Negative \n33 Negative Negative Negative \n34 Negative Negative Negative \n35 Positive Positive – Cq 14.84 Negative \n36 Negative Negative Negative \n37 Negative Negative Negative \n38 Positive Positive – Cq 21.81 Negative \n39 Negative Negative Negative \n40 Negative* Negative Negative \n41 Negative Negative Negative \n42 Negative Negative Negative \n43 Negative Negative Negative \n44 Negative Negative Negative \n45 Negative Negative Negative \n46 Negative Negative Negative \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n.CC-BY 4.0 International licenseperpetuity. It is made available under a \npreprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in \nThe copyright holder for thisthis version posted January 9, 2026. ; https://doi.org/10.64898/2026.01.09.698603doi: bioRxiv preprint \n\n \n17 \n \nReferences \n1. Ng TFF , Kondov NO, Deng X, Van Eenennaam A, Neibergs HL, Delwart E. A \nMetagenomics and Case-Control Study To Identify Viruses Associated with \nBovine Respiratory Disease. J Virol. 2015 May 15;89(10):5340–9.  \n2. Robinson E, Schulein C, Jacobson BT, Jones K, Sago J, Huber VC, et al. \nPathophysiology of Influenza D Virus Infection in Specific-Pathogen-Free Lambs \nwith or without Prior Mycoplasma ovipneumoniae Exposure. Viruses. 2022 Jul \n1;14(7).  \n3. Nedland H, Wollman J, Sreenivasan C, Quast M, Singrey A, Fawcett L, et al. \nSerological evidence for the co-circulation of two lineages of influenza D viruses \nin equine populations of the Midwest United States. Zoonoses Public Health. \n2018 Feb 1;65(1):e148–54.  \n4. Gorin S, Fablet C, Quéguiner S, Barbier N, Paboeuf F , Hervé S, et al. Assessment \nof influenza D virus in domestic pigs and wild boars in France: Apparent limited \nspread within swine populations despite serological evidence of breeding sow \nexposure. Viruses. 2019 Dec 24;12(1).  \n5. Salem E, Cook EAJ, Lbacha HA, Oliva J, Awoume F , Aplogan GL, et al. Serologic \nevidence for influenza c and d virus among ruminants and Camelids, Africa, \n1991-2015. Emerg Infect Dis. 2017 Sep 1;23(9):1556–9.  \n6. Sreenivasan C, Thomas M, Sheng Z, Hause BM, Collin EA, Knudsen DEB, et al. \nReplication and Transmission of the Novel Bovine Influenza D Virus in a Guinea \nPig Model. J Virol. 2015 Dec;89(23):11990–2001.  \n7. Song H, Qi J, Khedri Z, Diaz S, Yu H, Chen X, et al. An Open Receptor-Binding \nCavity of Hemagglutinin-Esterase-Fusion Glycoprotein from Newly-Identified \nInfluenza D Virus: Basis for Its Broad Cell Tropism. PLoS Pathog. 2016;12(1).  \n8. Nemanichvili N, Berends AJ, Wubbolts RW, Gröne A, Rijks JM, de Vries RP , et al. \nTissue microarrays to visualize influenza d attachment to host receptors in the \nrespiratory tract of farm animals. Viruses. 2021 Apr 1;13(4).  \n9. Holwerda M, Kelly J, Laloli L, Stürmer I, Portmann J, Stalder H, et al. Determining \nthe replication kinetics and cellular tropism of influenza D virus on primary well-\ndiƯ erentiated human airway epithelial cells. Viruses. 2019 Apr 1;11(4).  \n10. Borkenhagen LK, Mallinson KA, Tsao RW, Ha SJ, Lim WH, Toh TH, et al. \nSurveillance for respiratory and diarrheal pathogens at the human-pig interface in \nSarawak, Malaysia. PLoS One. 2018 Jul 1;13(7).  \n11. Hause BM, Ducatez M, Collin EA, Ran Z, Liu R, Sheng Z, et al. Isolation of a Novel \nSwine Influenza Virus from Oklahoma in 2011 Which Is Distantly Related to \nHuman Influenza C Viruses. PLoS Pathog. 2013 Feb;9(2).  \n12. Trombetta CM, Marchi S, Manini I, Kistner O, Li F , Piu P , et al. Influenza D virus: \nSerological evidence in the Italian population from 2005 to 2017. Viruses. 2019 \nDec 27;12(1).  \n13. White SK, Ma W, McDaniel CJ, Gray GC, Lednicky JA. Serologic evidence of \nexposure to influenza D virus among persons with occupational contact with \ncattle. Journal of Clinical Virology. 2016 Aug 1;81:31–3.  \n14. Ferguson L, Eckard L, Epperson WB, Long LP , Smith D, Huston C, et al. Influenza D \nvirus infection in Mississippi beef cattle. Virology. 2015 Dec 1;486:28–34.  \n15. Alvarez I, Carrera M, Chiapponi C, Ducatez M, El Agrebi N, Faccini S, et al. \nDeveloping an integrated approach to assess the emergence threat associated \n.CC-BY 4.0 International licenseperpetuity. It is made available under a \npreprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in \nThe copyright holder for thisthis version posted January 9, 2026. ; https://doi.org/10.64898/2026.01.09.698603doi: bioRxiv preprint \n\n \n18 \n \nwith influenza D viruses’ circulating in Europe. EFSA Supporting Publications. \n2024 Feb 6;21(2).  \n16. Baba MM, Bitew M, Fokam J, Lelo EA, Ahidjo A, Asmamaw K, et al. Diagnostic \nperformance of a colorimetric RT -LAMP for the identification of SARS-CoV-2: A \nmulticenter prospective clinical evaluation in sub-Saharan Africa. \nEClinicalMedicine. 2021 Oct 1;40.  \n17. Choi G, Moehling TJ, Meagher RJ. Advances in RT-LAMP for COVID-19 testing and \ndiagnosis. Vol. 23, Expert Review of Molecular Diagnostics. Taylor and Francis \nLtd.; 2023. p. 9–28.  \n18. Ducatez MF , Pelletier C, Meyer G. Influenza D Virus in Cattle, France, 2011–2014. \nEmerg Infect Dis [Internet]. 2015 Feb 4;21(2). Available from: \nhttp://wwwnc.cdc.gov/eid/article/21/2/14-1449_article.htm \n19. Salem E, Hägglund S, Cassard H, Corre T, Näslund K, Foret C, et al. Pathogenesis, \nHost Innate Immune Response, and Aerosol Transmission of Influenza D Virus in \nCattle. J Virol. 2019 Apr;93(7).  \n20. van Diemen PM, Ramsay AR, Bernard M, Floyd T, Byrne AMP , Khatri M, et al. \nInfluenza D virus in Great Britain [Internet]. 2023 [cited 2026 Jan 8]. Available \nfrom: \nhttps://www.researchgate.net/publication/388946074_Pig_Influenza_D_virus_in_\nGreat_Britain \n21. Faccini S, De Mattia A, Chiapponi C, Barbieri I, Boniotti MB, Rosignoli C, et al. \nDevelopment and evaluation of a new Real-Time RT-PCR assay for detection of \nproposed influenza D virus. J Virol Methods. 2017 May 1;243:31–4.  \n22. Lamas A, Azinheiro S, Roumani F , Prado M, Garrido-Maestu A. Evaluation of the \neƯ ect of outer primer structure, and inner primer linker sequences, in the \nperformance of Loop-mediated isothermal amplification. Talanta. 2023 Aug \n1;260.  \n  \n.CC-BY 4.0 International licenseperpetuity. It is made available under a \npreprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in \nThe copyright holder for thisthis version posted January 9, 2026. ; https://doi.org/10.64898/2026.01.09.698603doi: bioRxiv preprint","source_license":"CC-BY-4.0","license_restricted":false}