Colourimetric LAMP detection of Escherichia coli in bovine milk using a simplified DNA preparation method Authors: Tewodros Fentahun Jember, Mark Edward Westman, Sameer Dinkar Pant, Seyed Ali Ghorashi Published: 2026 ABSTRACT Escherichia coli (E. coli) is a common cause of bovine mastitis, and rapid detection from milk is important for timely disease management. However, molecular assays such as colourimetric loop-mediated isothermal amplification (LAMP) are often limited by laboratory-based DNA extraction. This study evaluated a simplified HotSHOT (HS) DNA preparation method for extracting E. coli DNA from bovine milk and assessed its compatibility with colourimetric LAMP and conventional PCR. Pasteurised bovine milk spiked with E. coli was used for analytical evaluation, and culture‑positive clinical milk samples were tested to assess diagnostic performance. The HS method was compared with five commercial DNA extraction kits and a direct boiling method without centrifugation. Performance was assessed using limit of detection (LOD) expressed as colony-forming units per millilitre. Using the HS method, the LOD was 7.9 × 10⁴ CFU/mL for LAMP and 7.9 × 10³ CFU/mL for PCR. PCR performance of the HS method was comparable to the QIAamp and Norgen commercial kits, while LAMP sensitivity was ten- to a hundred-fold lower than the best kit-based methods. Direct boiling showed markedly reduced sensitivity, with an LOD of 7.9 × 10⁸ CFU/mL. Among 17 culture-confirmed clinical mastitis milk samples positive for E. coli, LAMP detected 15 (88.2%) and PCR detected all 17 (100%). These results show that simplified DNA preparation can support molecular detection of E. coli in bovine milk and provide a practical basis for future development of field-deployable diagnostics in resource-constrained settings.Supplementary InformationThe online version contains supplementary material available at https://doi.org/10.1007/s11259-026-11576-8. Introduction Bovine mastitis remains a major health challenge in dairy production and leads to reduced milk yield, compromised animal welfare, and increased antimicrobial use (Ghumman et al. 2026, Ramuada et al. 2024; Hogeveen et al. 2011, Tomanic et al. 2025). Bacterial culture is considered the reference method for identifying mastitis-causing pathogens, but it is time-consuming and usually requires milk samples to be transported to diagnostic laboratories (Servellon et al. 2025; Ramuada et al. 2024). As a result, treatment decisions are often delayed or made before pathogen identification is available (Wilm et al. 2021; Preine et al. 2022). Several rapid cow-side tests, such as the California Mastitis Test and somatic cell count-based methods, are widely used to detect mastitis (Peckler and Adcock 2025; Ferronatto et al. 2018). While these tests are effective for identifying inflammation of the mammary gland, they do not identify the causative bacterial agent. This lack of pathogen-specific information often leads to non-specific treatment, which may reduce treatment success and contribute to unnecessary antimicrobial use (Saeed et al. 2024; Ajose et al. 2022). Molecular diagnostic methods offer sensitive and specific detection of mastitis causing bacteria directly from milk samples (Duarte et al. 2015; Kajdanek et al. 2024; Koskinen et al. 2009). Conventional polymerase chain reaction (PCR) has been widely adopted for this purpose, but it requires thermal cyclers, stable electricity, and trained personnel (Shome et al. 2011; Syring et al. 2012; Hiitio et al. 2015; Riffon et al. 2001). These requirements limit its use outside laboratory environments and restrict its application as a routine on-farm diagnostic tool (Wong et al. 2020; Poussard et al. 2025; Ghorashi et al. 2022; Garner et al. 2023). Loop-mediated isothermal amplification (LAMP) has emerged as a practical alternative to PCR for decentralised testing (Notomi et al. 2000). LAMP operates under isothermal conditions and does not require complex instrumentation (Cui et al. 2024; Garner et al. 2023; Notomi et al. 2000). The development of colourimetric LAMP assays has further simplified detection by allowing results to be interpreted visually through colour change, instead of relying on turbidity measurement, fluorescence detection, or gel electrophoresis, which require additional laboratory equipment (Tanner et al. 2015). These features make LAMP suitable for rapid testing in non-laboratory settings and as a potential cow side molecular assay (Kapalamula et al. 2021; Pascual-Garrigos et al. 2021). Among mastitis-causing pathogens, Escherichia coli (E. coli) is one of the most common environmental bacteria associated with both clinical and subclinical mastitis in dairy cattle (Goulart and Mellata 2022). Infection often arises from environmental contamination such as bedding, manure, and suboptimal housing conditions, and outbreaks are frequently linked to wet and unsanitary farm environments (Goulart and Mellata 2022). Rapid identification of E. coli is important because infections can progress quickly and may require management strategies that differ from those used for contagious mastitis pathogens (Goulart and Mellata 2022). Despite the advantages of molecular assays, their performance strongly depends on the quality of the extracted DNA (Ravine and Suthers 2012; Ghorashi et al. 2013). DNA extraction is often the most critical and limiting step in molecular diagnostics (Paul et al. 2020). Rapid DNA preparation methods are typically crude and may contain PCR inhibitors that reduce assay sensitivity (Date Chong et al. 2023). In contrast, commercial DNA extraction kits generally provide higher quality DNA but require laboratory equipment, multiple reagents, and higher costs (Silva et al. 2023). This mismatch between simple amplification methods (e.g. LAMP) and complex DNA extraction processes (e.g. column-based commercial kits) limits the broader adoption of molecular diagnostics outside laboratory settings (Zou et al. 2017). The aim of this study was to develop and evaluate a colourimetric LAMP assay for the detection of E. coli in bovine milk and to assess its compatibility with a simplified HotSHOT (HS) DNA preparation method. Both analytically controlled spiked milk samples and culture-confirmed clinical milk samples were used to assess assay performance. The performance of the HS method was compared with five commercial DNA extraction kits and a direct boiling method. Extracted DNA was tested using the developed LAMP assay and conventional PCR, and assay sensitivity and specificity were evaluated using the limit of detection expressed as colony-forming units per millilitre (CFU/mL). Materials and methods Preparation of E. coli–spiked milk samples An overview of the complete workflow, from milk sampling through HS DNA preparation to colourimetric LAMP and PCR detection, is shown in Fig. 1. A commercial pasteurised full-cream bovine milk was used to prepare E. coli–spiked samples for assay evaluation. E. coli colonies were obtained from the Veterinary Diagnostic Laboratory at Charles Sturt University and were originally isolated from bovine mastitis milk samples and identified as E. coli using selective differential media and standard biochemical tests. A loopful of colonies was inoculated into Luria-Bertani (LB) broth and cultured overnight at 37 °C with shaking. The identity of the isolate was confirmed by sequencing (NCBI accessions PV472575 to PV472578), and a characterised reference strain will be included in future work.Fig. 1Overview of the LAMP based workflow for detection of E. coli in bovine milk. Milk was sampled from the affected quarter and processed by a simplified HS DNA preparation. The prepared DNA was tested by colourimetric LAMP targeting the uidA gene, with results read visually as a colour change from pink to yellow. Conventional PCR, yielding a 207 bp product, was performed in parallel as a reference method for comparison Overview of the LAMP based workflow for detection of E. coli in bovine milk. Milk was sampled from the affected quarter and processed by a simplified HS DNA preparation. The prepared DNA was tested by colourimetric LAMP targeting the uidA gene, with results read visually as a colour change from pink to yellow. Conventional PCR, yielding a 207 bp product, was performed in parallel as a reference method for comparison Following incubation of the LB broth, 1 mL of the overnight culture was added to 9 mL of pasteurised full-cream milk in a sterile Falcon tube to prepare the initial dilution. A tenfold serial dilution series was then prepared from this suspension to generate dilutions ranging from 10⁻¹ to 10⁻¹⁰. All dilutions were prepared and tested in triplicate. LB agar was prepared using tryptone, yeast extract, and sodium chloride as previously described (MacWilliams and Liao 2006). The medium was autoclaved at 121 °C for 30 min, poured into Petri dishes, and allowed to solidify. From each dilution, 100 µL was spread onto LB agar plates. The agar plates were incubated at 37°C for 18 h for colony counting. After incubation, colonies were counted on plates containing between 30 and 300 colonies. Colony-forming units per mL were calculated based on the number of colonies observed, the volume plated, and the corresponding dilution factor. Each dilution was plated and counted directly, so the reported CFU per mL reflected the viable count in each tested tube rather than a value back calculated from the starting culture. DNA extracted from each serial dilution using the seven DNA extraction methods was used as a template for colourimetric LAMP and PCR assays. Results were expressed as limit of detection (LOD) to enable relative comparison of the seven DNA extraction methods. The LOD was defined as the lowest bacterial concentration, in CFU per mL, at which a positive result was obtained in at least two of three replicates for a given assay and extraction method. DNA extraction The optimised HS DNA preparation method consists of a lysis solution containing 0.2 mM disodium EDTA and 25 mM NaOH adjusted to pH 12.0, and a neutralisation solution containing 40 mM Tris HCl adjusted to pH 1.9. For each extraction, 150 µL of milk was mixed with 75 µL of lysis solution in a 1.5 mL tube, homogenised by pipetting and gentle finger flicking, and heated at 90 °C for 10 min on a heat block. After heating, 75 µL of neutralisation solution was added and mixed, giving a final volume of 300 µL. No centrifugation was required. DNA was extracted from each serial dilution of E. coli–spiked milk, ranging from 10⁻¹ to 10⁻¹⁰, using the optimised HS method and six additional DNA extraction methods listed in Table 1. Commercial kits were used according to the manufacturers’ instructions. For the direct boiling method, 150 µL of milk was heated at 100 °C for 10 min on a heat block, and the upper liquid phase was taken directly as template without centrifugation. Direct boiling without centrifugation was included because it is the simplest heat‑based extraction that could be performed on farm with only a heat block, matching the equipment requirements of the HS method. It therefore served as a heat‑only comparator for the alkaline lysis and neutralisation chemistry of HS. This procedure differs from conventional laboratory boiling protocols, in which bacterial cells are pelleted by centrifugation and resuspended in buffer before boiling. All extractions were performed using samples contained in sterile Falcon tubes. Table 1DNA extraction methods evaluated in this studyMethod IDExtraction methodSourceExtraction mechanism1Wizard® Genomic DNA Purification Kit (Gram Positive and Negative Bacteria) (Promega Corporation, Australia)Commercial kitEnzymatic, Mechanical2ReliaPrep™ Blood gDNA Miniprep System (Promega Corporation, Australia)Commercial kitEnzymatic, Mechanical3QIAmp DNA Microbiome Kit (QIAGEN, USA)Commercial kitMechanical, Enzymatic, Chemical4Milk Bacterial DNA Isolation Kit (Norgen Biotek Corporation, Ontario, Canada)Commercial kitMechanical, Enzymatic5Quick-DNA Fecal/Soil Microbe DNA Miniprep Kit (Zymo Research Corporation, Australia)Commercial kitMechanical, Enzymatic6Direct boiling method (no centrifugation)Non-commercialThermal7Optimised HS methodNon-commercial (test)Chemical, Thermal Table 1DNA extraction methods evaluated in this studyMethod IDExtraction methodSourceExtraction mechanism1Wizard® Genomic DNA Purification Kit (Gram Positive and Negative Bacteria) (Promega Corporation, Australia)Commercial kitEnzymatic, Mechanical2ReliaPrep™ Blood gDNA Miniprep System (Promega Corporation, Australia)Commercial kitEnzymatic, Mechanical3QIAmp DNA Microbiome Kit (QIAGEN, USA)Commercial kitMechanical, Enzymatic, Chemical4Milk Bacterial DNA Isolation Kit (Norgen Biotek Corporation, Ontario, Canada)Commercial kitMechanical, Enzymatic5Quick-DNA Fecal/Soil Microbe DNA Miniprep Kit (Zymo Research Corporation, Australia)Commercial kitMechanical, Enzymatic6Direct boiling method (no centrifugation)Non-commercialThermal7Optimised HS methodNon-commercial (test)Chemical, Thermal DNA extraction methods evaluated in this study Pasteurised full cream milk spiked with distilled water instead of overnight E. coli culture was used as a negative extraction control. The detailed extraction procedure for the optimised HS method has been described previously (Jember et al. 2026a). LAMP and PCR primer design Colourimetric LAMP primers were designed to detect E. coli by targeting the uidA gene, which encodes beta-glucuronidase. The target gene sequence was obtained from the NCBI GenBank database under accession number KT311783.1. The uidA gene, which encodes beta-glucuronidase, was selected because it is a widely accepted molecular marker specific to E. coli. Multiple candidate LAMP primer sets were generated in PrimerExplorer V5 (http://primerexplorer.jp/e/index.html) provided by Eiken Chemical Co. in Tokyo, Japan, and the final set was selected on the basis of predicted melting temperature, GC content, absence of strong primer dimers and hairpins, and location within a conserved region of the gene. The LAMP primer set consisted of an outer forward primer (EC-F3), an outer backward primer (EC-B3), a forward inner primer (EC-FIP), a backward inner primer (EC-BIP), a forward loop primer (EC-LF), and a backward loop primer (EC-LB). Primer binding sites were mapped relative to the uidA gene sequence of E. coli (GenBank accession KT311783.1). Nucleotide positions were determined for the outer and loop primers, while the inner primers FIP and BIP were designed as composite primers derived from internal regions of the target gene, consistent with standard LAMP architecture. The EC F3 and EC B3 primers were also used as forward and reverse primers for conventional PCR amplification, generating an expected amplicon size of 207 base pairs. All primers were synthesised by Sigma-Aldrich Pty Ltd in Australia. Primer sequences are listed in Table 2, and primer binding locations within the target sequence are illustrated in Fig. 2. Primer sequences are colour coded in Table 2 to allow clear visual correspondence with their binding locations shown in Fig. 2.Fig. 2Schematic representation of LAMP primer binding sites within the uidA gene of E. coli (GenBank accession KT311783.1). Each primer is colour coded to indicate its specific binding region. The locations of outer primers (F3 and B3) and loop primers (LF and LB) are shown with nucleotide coordinates relative to the reference sequence. Inner primers FIP and BIP are composite primers comprising two target specific regions and are therefore not represented by single continuous binding positionsFig. 3Determination of LOD for LAMP and PCR using E. coli spiked milk samples. (A) Colourimetric LAMP results obtained from tenfold serial dilutions of full cream milk spiked with E. coli. Labels 1 to 10 correspond to dilutions 10⁻¹ to 10⁻¹⁰. Each dilution was tested in triplicate. A colour change from pink to yellow was interpreted as a positive result. Boxed tubes indicate the dilution at which the LOD was determined for each DNA extraction method. N indicates negative control and P indicates positive control. (B) Agarose gel electrophoresis of PCR amplicons generated using DNA extracted from the same serially diluted milk samples. Lanes 1 to 10 correspond to dilutions 10⁻¹ to 10⁻¹⁰. M indicates DNA ladder, N negative control, and P positive control. PCR products were separated on 1.5% agarose gels stained with GelRed. LOD values for each DNA extraction method are expressed as CFU/mL. The boxed tube in (A) and the last visible band in (B) indicate the limit of detection for each method; values are given in Table 3. DNA extraction methods are abbreviated as follows: RL, ReliaPrep Blood gDNA Miniprep System; QI, QIAamp DNA Microbiome Kit; BL, direct boiling method; NR, Milk Bacterial DNA Isolation Kit; HS, optimised HS method; ZY, Quick DNA Fecal Soil Microbe DNA Miniprep Kit; WZ, Wizard Genomic DNA Purification Kit Schematic representation of LAMP primer binding sites within the uidA gene of E. coli (GenBank accession KT311783.1). Each primer is colour coded to indicate its specific binding region. The locations of outer primers (F3 and B3) and loop primers (LF and LB) are shown with nucleotide coordinates relative to the reference sequence. Inner primers FIP and BIP are composite primers comprising two target specific regions and are therefore not represented by single continuous binding positions Determination of LOD for LAMP and PCR using E. coli spiked milk samples. (A) Colourimetric LAMP results obtained from tenfold serial dilutions of full cream milk spiked with E. coli. Labels 1 to 10 correspond to dilutions 10⁻¹ to 10⁻¹⁰. Each dilution was tested in triplicate. A colour change from pink to yellow was interpreted as a positive result. Boxed tubes indicate the dilution at which the LOD was determined for each DNA extraction method. N indicates negative control and P indicates positive control. (B) Agarose gel electrophoresis of PCR amplicons generated using DNA extracted from the same serially diluted milk samples. Lanes 1 to 10 correspond to dilutions 10⁻¹ to 10⁻¹⁰. M indicates DNA ladder, N negative control, and P positive control. PCR products were separated on 1.5% agarose gels stained with GelRed. LOD values for each DNA extraction method are expressed as CFU/mL. The boxed tube in (A) and the last visible band in (B) indicate the limit of detection for each method; values are given in Table 3. DNA extraction methods are abbreviated as follows: RL, ReliaPrep Blood gDNA Miniprep System; QI, QIAamp DNA Microbiome Kit; BL, direct boiling method; NR, Milk Bacterial DNA Isolation Kit; HS, optimised HS method; ZY, Quick DNA Fecal Soil Microbe DNA Miniprep Kit; WZ, Wizard Genomic DNA Purification Kit Table 2Sequences and positions of primers targeting the uidA gene of E. coli used for LAMP and PCRPrimer namePrimer Sequence (5’−3’)Position of primers in uidA geneaEC-F3CGCAGCGTAATGCTCTACAC198–217EC-B3GTGTGGATTCACCACTTGCA385–404EC-FIPAGACGCGTGGTTACAGTCTTGC-GAACACCTGGGTGGACGAF2: 224–241 and F1c: 266–287EC-BIPTGGTGGCCAATGGTGATGTCA-TCCCGCTAGTGCCTTGTCB1c: 301–321 and B2: 364–381EC-LFTGCGTCACCACGGTGATA242–259EC-LBGCGTTGAACTGCGTGATGC322–340a Positions are given relative to the uidA gene sequence of E. coli (GenBank accession KT311783.1) Table 2Sequences and positions of primers targeting the uidA gene of E. coli used for LAMP and PCRPrimer namePrimer Sequence (5’−3’)Position of primers in uidA geneaEC-F3CGCAGCGTAATGCTCTACAC198–217EC-B3GTGTGGATTCACCACTTGCA385–404EC-FIPAGACGCGTGGTTACAGTCTTGC-GAACACCTGGGTGGACGAF2: 224–241 and F1c: 266–287EC-BIPTGGTGGCCAATGGTGATGTCA-TCCCGCTAGTGCCTTGTCB1c: 301–321 and B2: 364–381EC-LFTGCGTCACCACGGTGATA242–259EC-LBGCGTTGAACTGCGTGATGC322–340a Positions are given relative to the uidA gene sequence of E. coli (GenBank accession KT311783.1) Sequences and positions of primers targeting the uidA gene of E. coli used for LAMP and PCR a Positions are given relative to the uidA gene sequence of E. coli (GenBank accession KT311783.1) Colourimetric LAMP assay The colourimetric LAMP assay was initially developed and optimised using purified E. coli genomic DNA extracted from the same bacterial isolate described above using the Wizard SV Genomic DNA Purification System (Promega). The assay was performed using a modified WarmStart® Colourimetric LAMP 2× Master Mix (New England Biolabs, USA). Each reaction was carried out in a final volume of 20 µL containing 10 µL of 2× LAMP master mix, primers at final concentrations of 0.2 µM each for EC-F3 and EC-B3, 0.4 µM each for EC-LF and EC-LB, and 1.6 µM each for EC-FIP and EC-BIP, 5 µL of nuclease-free water, and 3 µL of extracted DNA. The LAMP and PCR reactions used 3 µL and 2 µL of template in total volumes of 20 µL and 25 µL respectively, so the proportion of template was similar, and the same extracted DNA was used for both assays. For negative controls, both nuclease free water and DNA extracted from non-spiked pasteurised milk were used. For positive controls, 2 µL of E. coli genomic DNA at 1 ng/µL, extracted using the same commercial kit, was included. LAMP reactions were incubated at 65 °C on a heat block for 60 min. The relatively long incubation reflects amplification from a crude milk extract, and the reduced speed advantage of LAMP under these conditions is addressed in the discussion. For DNA extracted using the optimised HS method, the incubation time was extended by an additional 30min. This adjustment was made to allow sufficient amplification from a crude DNA preparation that may contain lower DNA concentrations or residual inhibitors compared to purified kit-based extractions. This extended incubation time was applied consistently for all experiments using HS-extracted DNA. According to the manufacturer’s instructions, a colour change from pink to yellow was interpreted as a positive result, while no colour change was considered negative. Conventional PCR assay Conventional PCR was performed using a Rotor Gene™ 6000 thermal cycler (Qiagen, Melbourne, Australia). Each reaction was carried out in a final volume of 25 µL containing 5.0 µL of 5× GoTaq Green Flexi Reaction Buffer, 1.5 µL of MgCl₂ (25 mM), 2.0 µL of dNTP mix (1.25 mM each), 0.3 µL of Taq DNA polymerase (5 U/µL; Promega, Australia), 1.0 µL each of EC-F3 and EC-B3 primers (2 µM), and 2.0 µL of extracted DNA. Nuclease free water was added to reach the final reaction volume. Negative and positive controls were prepared as described for the colourimetric LAMP assay. PCR amplification was performed with an initial denaturation at 95 °C for 5 min, followed by 30 cycles of denaturation at 95 °C for 1 min, annealing at 56 °C for 45 s, and extension at 72 °C for 70 s, with a final extension at 72 °C for 8 min. The PCR products (10 µL) were analysed by agarose gel electrophoresis on a 1.5% agarose gel and visualised using a Gel Doc XR+ imaging system (Gel Doc XR+, Bio-Rad, USA). Limit of detection (analytical sensitivity) The performance of different DNA extraction methods was compared based on the limit of detection (LOD) for both colourimetric LAMP and PCR assays. For each serial dilution, bacterial concentration was determined by colony counting and expressed as CFU/mL. The LOD was defined as the lowest bacterial concentration at which a positive result was obtained in at least two of three replicates for a given assay and DNA extraction method. Where minor variation in LOD was observed across the three technical replicates for a given extraction method, the value obtained in the majority of the replicates was taken as the representative LOD (Supplementary Fig. 1). Diagnostic sensitivity assessment using clinical milk samples Following the comparative evaluation of DNA extraction methods using spiked milk samples, the diagnostic sensitivity of the optimised HS method was assessed using clinical milk samples obtained from the Elizabeth Macarthur Agricultural Institute in Australia. A total of 17 clinical bovine milk samples that were culture positive for E. coli were included. Identification of E. coli was confirmed by matrix-assisted laser desorption ionisation time-of-flight mass spectrometry. DNA was extracted from these samples using the optimised HS method and tested using colourimetric LAMP and conventional PCR assays. Diagnostic sensitivity and specificity were calculated using a 2× 2 contingency table, with bacterial culture as the reference method. Calculations were performed using the MedCalc online diagnostic test evaluation tool (https://www.medcalc.org/calc/diagnostic_test.php). Specificity assessment The specificity of the colourimetric LAMP and PCR assays were evaluated by testing the primers against a panel of non-target bacterial species selected to assess potential cross reactivity. The panel included Streptococcus uberis, Pasteurella multocida, Mycoplasma gallisepticum, Erysipelothrix species, Staphylococcus aureus, Mycoplasma synoviae, Streptococcus zooepidemicus, multidrug resistant Staphylococcus aureus, Pseudomonas aeruginosa, Enterococcus faecalis, and Salmonella Typhimurium. Genomic DNA from E. coli was used as a positive control, while distilled water was included as a negative control. In addition, the specificity of the outer primers EC‑F3 and EC‑B3 was assessed in silico by BLASTn against the NCBI core nucleotide database (accessed September 2026), with the maximum number of target sequences set to 500. Searches were performed without restriction and then restricted in turn to Shigella, Klebsiella pneumoniae, Enterobacter cloacae, Citrobacter and Serratia marcescens. A full match was defined as 100% identity across the whole primer length. Statistical analysis The representative limit of detection for each method and assay was taken as the value obtained in the majority of the three replicates. No formal statistical test of reproducibility was applied, because the replicates were technical rather than independent. Results Development and optimisation of colourimetric LAMP and PCR assays Under the optimised conditions, the colourimetric LAMP assay produced a clear colour change from pink to yellow in the presence of E. coli DNA, while no colour change was observed in negative control reactions containing nuclease-free water. Similarly, conventional PCR amplification using EC-F3 and EC-B3 primers consistently generated the expected 207 bp amplicon from E. coli genomic DNA, with no amplification detected in negative control reactions. These results confirmed that both assays were functional and suitable for subsequent analytical evaluation. Comparison of DNA extraction methods and assay limits of detection using E. coli spiked pasteurised bovine milk The limits of detection agreed to within one dilution across the three technical replicates for each extraction method and assay (Table 3; Supplementary Fig. 1). Using the optimised HS method, colourimetric LAMP detected E. coli at concentrations of approximately 7.9 × 10⁴ CFU/mL, while PCR detected E. coli at a representative LOD of approximately 7.9 × 10³ CFU/mL. The QIAamp DNA Microbiome Kit and the Milk Bacterial DNA Isolation Kit gave PCR limits of detection comparable to HS (7.9 × 10³ CFU/mL), but tenfold lower LAMP limits (7.9 × 10³ CFU/mL). Table 3LOD of E. coli determined by LAMP and PCR across three replicate samples using different DNA extraction methodsDNA extraction methodLAMP LOD (CFU/mL)PCR LOD (CFU/mL)E1E2E3E1E2E3RL7.9 × 10²7.9 × 10³7.9 × 10²7.9 × 10¹7.9 × 10²7.9 × 10¹QI7.9 × 10³7.9 × 10³7.9 × 10³7.9 × 10³7.9 × 10³7.9 × 10²NR7.9 × 10³7.9 × 10³7.9 × 10³7.9 × 10³7.9 × 10³7.9 × 10³HS7.9 × 10⁴7.9 × 10⁴7.9 × 10⁴7.9 × 10³7.9 × 10⁴7.9 × 10³ZY7.9 × 10⁵7.9 × 10⁵7.9 × 10⁴7.9 × 10³7.9 × 10⁴7.9 × 10⁴WZ7.9 × 10²7.9 × 10²7.9 × 10²7.9 × 10¹7.9 × 10¹7.9 × 10²BL7.9 × 10⁸7.9 × 10⁸7.9 × 10⁷7.9 × 10⁸7.9 × 10⁷7.9 × 10⁸Values represent limits of detection expressed as CFU/mL. E1, E2, and E3 indicate three replicate samples tested from a single DNA extraction. RL: ReliaPrep Blood gDNA Miniprep System (Promega Corporation, Australia); NR: Milk Bacterial DNA Isolation Kit (Norgen Biotek Corporation, Canada); ZY: Quick DNA Fecal Soil Microbe DNA Miniprep Kit (Zymo Research Corporation, Australia); QI: QIAamp DNA Microbiome Kit (QIAGEN, USA); HS: optimised HotSHOT method; WZ: Wizard Genomic DNA Purification Kit (Promega Corporation, Australia); BL: direct boiling method. The limit of detection agreed across the three replicates to within one dilution, and the representative value was that obtained in the majority of the three replicates Table 3LOD of E. coli determined by LAMP and PCR across three replicate samples using different DNA extraction methodsDNA extraction methodLAMP LOD (CFU/mL)PCR LOD (CFU/mL)E1E2E3E1E2E3RL7.9 × 10²7.9 × 10³7.9 × 10²7.9 × 10¹7.9 × 10²7.9 × 10¹QI7.9 × 10³7.9 × 10³7.9 × 10³7.9 × 10³7.9 × 10³7.9 × 10²NR7.9 × 10³7.9 × 10³7.9 × 10³7.9 × 10³7.9 × 10³7.9 × 10³HS7.9 × 10⁴7.9 × 10⁴7.9 × 10⁴7.9 × 10³7.9 × 10⁴7.9 × 10³ZY7.9 × 10⁵7.9 × 10⁵7.9 × 10⁴7.9 × 10³7.9 × 10⁴7.9 × 10⁴WZ7.9 × 10²7.9 × 10²7.9 × 10²7.9 × 10¹7.9 × 10¹7.9 × 10²BL7.9 × 10⁸7.9 × 10⁸7.9 × 10⁷7.9 × 10⁸7.9 × 10⁷7.9 × 10⁸Values represent limits of detection expressed as CFU/mL. E1, E2, and E3 indicate three replicate samples tested from a single DNA extraction. RL: ReliaPrep Blood gDNA Miniprep System (Promega Corporation, Australia); NR: Milk Bacterial DNA Isolation Kit (Norgen Biotek Corporation, Canada); ZY: Quick DNA Fecal Soil Microbe DNA Miniprep Kit (Zymo Research Corporation, Australia); QI: QIAamp DNA Microbiome Kit (QIAGEN, USA); HS: optimised HotSHOT method; WZ: Wizard Genomic DNA Purification Kit (Promega Corporation, Australia); BL: direct boiling method. The limit of detection agreed across the three replicates to within one dilution, and the representative value was that obtained in the majority of the three replicates LOD of E. coli determined by LAMP and PCR across three replicate samples using different DNA extraction methods Values represent limits of detection expressed as CFU/mL. E1, E2, and E3 indicate three replicate samples tested from a single DNA extraction. RL: ReliaPrep Blood gDNA Miniprep System (Promega Corporation, Australia); NR: Milk Bacterial DNA Isolation Kit (Norgen Biotek Corporation, Canada); ZY: Quick DNA Fecal Soil Microbe DNA Miniprep Kit (Zymo Research Corporation, Australia); QI: QIAamp DNA Microbiome Kit (QIAGEN, USA); HS: optimised HotSHOT method; WZ: Wizard Genomic DNA Purification Kit (Promega Corporation, Australia); BL: direct boiling method. The limit of detection agreed across the three replicates to within one dilution, and the representative value was that obtained in the majority of the three replicates The Quick DNA Fecal/Soil Microbe DNA Miniprep Kit showed higher detection limits, with LOD values of approximately 7.9 × 10⁵ CFU/mL for colourimetric LAMP and 7.9 × 10⁴ CFU/mL for PCR. The ReliaPrep Blood gDNA Miniprep System and the Wizard Genomic DNA Purification Kit demonstrated the highest sensitivity among the commercial kits, with limits of detection of 7.9 × 10² CFU/mL for LAMP and 7.9 × 10¹ CFU/mL for PCR. In contrast, the direct boiling method gave an LOD of 7.9 × 10⁸ CFU/mL for both assays, several orders of magnitude higher than that of the HS and kit-based methods. A detailed summary of the limits of detection obtained for each DNA extraction method using colourimetric LAMP and PCR is presented in Table 3, and a graphical comparison of detection limits is shown in Fig. 3. Clinical validation and diagnostic sensitivity of the colourimetric LAMP and PCR assays The diagnostic sensitivity of the colourimetric LAMP and PCR assays was assessed using 17 clinical bovine milk samples that were culture positive for E. coli. DNA was extracted from all samples using the optimised HS method and tested by colourimetric LAMP and conventional PCR. Among the 17 clinical milk samples tested, 15 were positive by colourimetric LAMP, corresponding to a diagnostic sensitivity of 88.24% (95% CI: 0.63–0.98). All 17 samples were positive by PCR, indicating a diagnostic sensitivity of 100%. No false positive reactions were observed in negative control samples for either assay. Representative colourimetric LAMP results and corresponding PCR amplification profiles for the clinical milk samples are shown in Fig. 4.Fig. 4Clinical validation of the LAMP and PCR assays for detection of E. coli in bovine milk samples. (A) Colourimetric LAMP results for 17 clinical milk samples (tubes 1–17) using template DNA extracted with the optimised HS method. N indicates negative control and P indicates positive control. (B) Agarose gel electrophoresis of PCR amplicons generated from the same clinical milk samples (n = 17) targeting the uidA gene of E. coli. Template DNA was extracted using the optimised HS method. PCR products were separated on 1.5% agarose gel stained with GelRed and visualised under UV illumination. M indicates DNA ladder, N negative control, and P positive control Clinical validation of the LAMP and PCR assays for detection of E. coli in bovine milk samples. (A) Colourimetric LAMP results for 17 clinical milk samples (tubes 1–17) using template DNA extracted with the optimised HS method. N indicates negative control and P indicates positive control. (B) Agarose gel electrophoresis of PCR amplicons generated from the same clinical milk samples (n = 17) targeting the uidA gene of E. coli. Template DNA was extracted using the optimised HS method. PCR products were separated on 1.5% agarose gel stained with GelRed and visualised under UV illumination. M indicates DNA ladder, N negative control, and P positive control Specificity of the colourimetric LAMP and PCR assays The colourimetric LAMP and PCR assays showed no cross-reactivity with any of the 11 non-target bacterial species tested (Table 4; Fig. 5). Only E. coli produced a positive result in both assays. Sequencing of selected PCR amplicons confirmed their identity as E. coli, and the sequences were deposited in NCBI under accession numbers PV472575, PV472576, PV472577, and PV472578.Table 4Analytical specificity of the LAMP and PCR assays against non-target bacterial speciesBacterial species testedLAMP resultPCR resultStreptococcus uberis--Pasteurella multocida--Mycoplasma gallisepticum--Erysipelothrix spp.--Staphylococcus aureus--Mycoplasma synoviae--Streptococcus zooepidemicus--Methicillin-resistant Staphylococcus aureus--Pseudomonas aeruginosa--Enterococcus faecalis--Salmonella Typhimurium--Escherichia coli++Negative control--“+” = positive; “-” = negative, Two isolates of S. aureus were tested: one methicillin-sensitive and one methicillin-resistantFig. 5Specificity assessment of the colourimetric LAMP and PCR assays. PCR products were analysed on a 1.5% agarose gel. The expected amplicon size was 207 bp (A) Colourimetric LAMP results and (B) PCR amplification using primers designed for E. coli, M indicates the DNA ladder, lanes 1–11 correspond to non-target bacterial isolates, lane 12 corresponds to E. coli and lane 13 represents the negative control with no template DNA Table 4Analytical specificity of the LAMP and PCR assays against non-target bacterial speciesBacterial species testedLAMP resultPCR resultStreptococcus uberis--Pasteurella multocida--Mycoplasma gallisepticum--Erysipelothrix spp.--Staphylococcus aureus--Mycoplasma synoviae--Streptococcus zooepidemicus--Methicillin-resistant Staphylococcus aureus--Pseudomonas aeruginosa--Enterococcus faecalis--Salmonella Typhimurium--Escherichia coli++Negative control--“+” = positive; “-” = negative, Two isolates of S. aureus were tested: one methicillin-sensitive and one methicillin-resistant Analytical specificity of the LAMP and PCR assays against non-target bacterial species “+” = positive; “-” = negative, Two isolates of S. aureus were tested: one methicillin-sensitive and one methicillin-resistant Specificity assessment of the colourimetric LAMP and PCR assays. PCR products were analysed on a 1.5% agarose gel. The expected amplicon size was 207 bp (A) Colourimetric LAMP results and (B) PCR amplification using primers designed for E. coli, M indicates the DNA ladder, lanes 1–11 correspond to non-target bacterial isolates, lane 12 corresponds to E. coli and lane 13 represents the negative control with no template DNA In silico analysis showed that both outer primers matched E. coli sequences with 100% identity and full query coverage. Full matches were also found for Shigella species. No full-length match was found for K. pneumoniae, E. cloacae, Citrobacter species or S. marcescens, for which only partial alignments were obtained. Discussion The optimised HS method produced DNA of sufficient quality for both LAMP and PCR detection of E. coli in bovine milk, with LOD comparable to several commercial extraction kits. These results demonstrate that a colour-based LAMP assay combined with a simplified DNA preparation method can provide a practical approach for detecting this pathogen. This supports the concept that complex and resource-intensive DNA extraction procedures are not always necessary for reliable molecular detection, particularly for decentralised or near-farm testing applications (Shwani et al. 2023). The LAMP chemistry, the uidA target, and the alkaline lysis approach are each established and are not presented as novel in themselves. The contribution of this work is the integration of a crude DNA preparation performed directly on milk with a visual colourimetric readout, benchmarked against five commercial kits. A key outcome of this work was the consistent performance of the HS method, with LOD values agreeing to within one dilution across technical replicates. Repeatable performance is a critical requirement for on-farm diagnostics, where testing conditions may vary (da Silva et al. 2021). In contrast, direct boiling of whole milk without centrifugation gave an LOD of 7.9 × 10⁸ CFU/mL. This method was included as the simplest field-compatible alternative to HS, requiring the same equipment and no centrifugation. Its poor performance indicates that heat alone does not release amplifiable DNA from milk, and that the alkaline lysis and neutralisation steps account for the performance of HS. The high LOD is likely to reflect carryover of milk fat, proteins and divalent cations that inhibit amplification (Habib et al. 2024; Chen et al. 2019), rather than an inherent limitation of boiling. Conventional boiling protocols that pellet and wash bacterial cells would be expected to perform considerably better (Cui et al. 2024), but they add a centrifugation step that the HS method avoids. The LAMP assay developed in this study demonstrated analytical sensitivity within the same order of magnitude as PCR when tested using E. coli spiked pasteurised bovine milk samples and DNA extracted using the optimised HS method (LAMP: 7.9 × 10⁴ CFU/mL; PCR: 7.9 × 10³ CFU/mL). Although PCR consistently achieved lower detection limits, both assays detected E. coli within a similar range, with PCR reaching the 10³ CFU/mL range and LAMP the 10⁴ CFU/mL range, indicating that the HS method provided DNA of sufficient quality for reliable detection by either method. This finding aligns with previous studies reporting that LAMP is a robust amplification method capable of tolerating certain inhibitors and operating effectively under isothermal conditions (Jackson et al. 2020). However, the results also confirm that LAMP performance remains dependent on DNA quality, reinforcing that DNA preparation is still a critical step even for isothermal amplification assays (Boza et al. 2024; Nwe et al. 2024; Moehling et al. 2021). The LAMP incubation in this study was 60min, and together with the 30-minute HS DNA preparation, the total time to a result was about 90 min, which is longer than the fastest reported LAMP run times and reduces its usual speed advantage. This reflects amplification from a crude milk extract, and optimising the enzyme, master mix, and reaction conditions is a clear priority for shortening the assay in future work. The assay did not reach the analytical sensitivity typically reported for real‑time PCR, which was not evaluated directly in this study, and its total time to result was similar to a laboratory qPCR run. Its intended value therefore lies not in replacing qPCR but in where it can be performed. Point‑of‑care diagnostics are generally judged on whether they are fit for purpose in their intended setting, accepting some loss of analytical sensitivity in exchange for affordability, simplicity and independence from laboratory equipment (Land et al. 2019). Real‑time PCR requires an optical thermal cycler and trained personnel to interpret amplification curves, and in routine mastitis diagnostics milk samples are sent to a diagnostic laboratory, with results typically available on the same or the following day (Hiitiö et al. 2015). Portable real‑time PCR instruments are available, but they remain considerably more expensive than the heat block required for the HS and LAMP workflow. In contrast, the HS method costs approximately AU$0.50 per sample, compared with AU$6 to AU$22 for commercial extraction kits (Jember et al. 2026b), and together with colourimetric LAMP gives a visual colour change that can be read without specialised training, although borderline colour changes may still require confirmation. Because the LOD obtained here lies below bacterial counts associated with clinical infection (Haider et al. 2023), the assay is best suited to the rapid identification of E. coli in clinical mastitis. However, LAMP did not detect 2 of 17 culture‑positive clinical samples, and laboratory qPCR remains the more appropriate choice for low bacterial loads and subclinical infection. E. coli was selected as the target organism in this study due to its relevance as a common environmental cause of bovine mastitis and its association with rapid disease progression, making it a suitable model pathogen for evaluating field-deployable molecular diagnostics. The bacterial load in milk, expressed as CFU/mL, is an important indicator of udder infection and mastitis status. In this study, the LOD of the LAMP and PCR assays using DNA extracted by the optimised HS method were 7.9 × 104 and 7.9 × 10³ CFU/mL, respectively. These detection limits are below the threshold associated with active mastitis, as bacterial counts above 1 × 10⁵ CFU/mL have been reported as indicative of infection in dairy cattle (Haider et al. 2023). Although lower LOD values for LAMP-based detection of E. coli have been reported elsewhere, such as the 5.7 × 10² CFU/mL reported by Cui et al. (2024), those studies relied on laboratory-adapted boiling protocols involving bacterial culture and multiple centrifugation steps. In contrast, DNA extraction in the present study was performed directly from milk, a complex biological matrix containing fat, proteins, and other components known to interfere with molecular amplification. Under these conditions, the HS method represents a compromise between analytical sensitivity and operational simplicity, supporting its use in resource constrained laboratory settings and near farm applications. When applied to clinical milk samples, the LAMP assay detected E. coli in 15 of 17 culture-positive samples, with a diagnostic sensitivity of 88.24%, whereas PCR detected E. coli in all 17 samples tested. The reduced sensitivity observed for LAMP in clinical samples may be attributed to lower bacterial loads, heterogeneous distribution of bacteria in milk, or the presence of inhibitory components that affect colour development or amplification efficiency (Nwe et al. 2024). These factors are well recognised challenges in direct milk testing and highlight the trade-off between assay simplicity and analytical sensitivity in field-deployable diagnostics (Kase and Pfefer 2016, Nwe et al. 2024). High specificity was observed for both LAMP and PCR, with no cross-reactivity detected against non-target bacterial species. This indicates that primers targeting the uidA gene provide specific detection of E. coli, which is essential for mastitis diagnostics where milk samples may contain a diverse microbial population. In silico analysis supported this, since the primers did not fully match the closely related coliform mastitis pathogens Klebsiella, Enterobacter, Citrobacter or Serratia. They did match Shigella, which is genetically almost indistinguishable from E. coli and is not a cause of bovine mastitis, so this has no practical consequence for the intended use of the assay. Accurate pathogen identification is critical for guiding appropriate treatment and avoiding unnecessary antimicrobial use (Yamin et al. 2023; Bhattacharya 2013; Ku et al. 2023). Overall, this study demonstrates that combining a colour-based LAMP assay with the HS DNA preparation method provides a balanced solution between analytical performance and operational simplicity. While PCR remains more sensitive, the HS-LAMP workflow reduces reliance on laboratory infrastructure, specialised equipment, and trained personnel, supporting its potential application as a near-farm or on-farm diagnostic tool (Bath et al. 2020; Lai et al. 2023; Lee et al. 2019; Yang et al. 2022; Sowmya et al. 2012). Despite the promising results obtained in this study, several limitations should be acknowledged. The number of clinical milk samples used for assay validation was relatively small, which limits the precision of the estimated diagnostic sensitivity and may not fully capture the variability present under field conditions. In addition, clinical samples were pre-characterised as E. coli positive by bacterial culture, and inclusion of a larger and more diverse sample set, including culture-negative samples and samples containing mixed infections, would allow more comprehensive evaluation of diagnostic specificity. The assay was evaluated using laboratory-prepared spiked milk samples and DNA extracted under controlled conditions, and performance may vary when applied directly on farm, where environmental factors, sample handling, and operator variability could influence results. In addition, the boiling comparator did not include the centrifugation and resuspension steps used in conventional boiling protocols, so its results should not be taken as representative of boiling extraction in general. Furthermore, although the colour-based readout simplifies result interpretation, borderline colour changes may introduce subjectivity, particularly at low bacterial concentrations. We also did not assess the purity of the HS extracts spectrophotometrically, since crude alkaline preparations are not well suited to standard purity ratios; the preparation was instead judged functionally by its ability to support reliable amplification. Future work should focus on larger-scale clinical validation and evaluation under true field conditions to further assess assay robustness, usability, and impact on mastitis management practices (Jones et al. 2019). Supplementary Information Below is the link to the electronic supplementary material. Supplementary Material 1 (PNG 139 KB) Supplementary Material 1 (PNG 139 KB)