Validation of mitochondrial COX3 primers for quantitative PCR-based authentication of chicken-derived products Authors: Saradiva Bilqisyah Azka, Farouq Heidar Barido, Slamet Diah Volkandari, Dicky Tri Utama, Yuny Erwanto, Muhammad Cahyadi Published: 2026 ABSTRACT AbstractThis study developed and validated a novel, highly specific primer set targeting the mitochondrial cytochrome c oxidase subunit III (COX3) gene for the detection of chicken DNA (Gallus gallus domesticus) using quantitative PCR (qPCR) to enhance food authentication reliability. A 70 bp chicken-specific amplicon was designed and validated for repeatability, analytical specificity across six animal species, and sensitivity. The assay demonstrated high precision, with an average quantification cycle (Cq) of 13.18 ± 0.20 and a stable melting temperature (Tm) of 81.27 ± 0.05 °C (RSDr <1.5%). Analytical specificity tests confirmed the detection of chicken DNA with no cross-amplification in non-target species, including beef, pork, sheep, squid, shrimp, and ducks. The absolute functional limit of detection (LoD) was 10.09 pg/μL, while the relative LoD in binary DNA mixtures reached 0.5% (v/v) with 100% PCR efficiency and R2 = 0.999. Application of this assay to 17 commercial meat products (sausages and floss) revealed a high incidence of mislabeling, where undeclared chicken DNA was detected in seven out of 11 samples (63.6%) declared as 100% beef, representing 41.2% of the total samples tested. These findings indicate that the COX3 primer set provides an accurate and efficient approach for authenticating animal-derived ingredients, significantly enhancing consumer protection against food adulteration.Keywords: Chicken DNA detection, COX3, Quantitative PCR, Primer design, Food authentication Introduction Food authenticity is a fundamental pillar of consumer protection and integrity. It ideally includes a detailed, verifiable, and accurate representation of the product's identity, composition, origin, and production claims (Ab-Latif et al., 2024). However, economically motivated adulteration (EMA), including undeclared species substitution in meat products, has emerged as a pervasive global threat to essential aspects of fair trade, market economics, public health, religious dietary compliance, and vulnerable consumers with species-specific allergies (Barido et al., 2023). The Codex Alimentarius Commission (CAC/GL 77-2017) explicitly prohibits food fraud as “intentional misleading of purchasers regarding the nature, substance, or quality of food for economic gain.” In addition, the harmonized regulations of European Union Regulation (EU) No. 1169/2011 mandate that food information must not mislead consumers concerning characteristics, including species composition, origin, or production method, with non-compliance constituting a criminal offense (Pádua et al., 2019). Despite these stringent international standards, the inherent complexity of modern, globalized meat supply chains facilitates opportunities for meat adulteration. Empirical evidence confirms the prevalence of such practices, including the 2013 European horse meat scandal, which revealed undeclared equine DNA (>1% w/w) in over 4.5 million tons of beef-labeled products across 16 countries, exposing systemic labeling failures (Premanandh, 2013). In addition, a 20-year systematic review of the global beef supply chain documented over 1200 EMA incidents, with species substitution comprising 42% of the verified cases (Robson et al., 2020). Furthermore, Indonesian market surveys have detected porcine and poultry DNA in beef-labeled meatballs, which is inconsistent with halal labeling and violates religious and regulatory standards (Setiani et al., 2025; Volkandari et al., 2025; Wibowo et al., 2023). Quantitative real-time PCR (qPCR) targeting mitochondrial DNA (mtDNA) loci has emerged as the analytical gold standard for species authentication in meat products, offering greater specificity and sensitivity than physicochemical techniques such as near-infrared (NIR) spectrophotometry, particularly in complex food matrices at trace concentrations (Volkandari et al., 2025). This is especially true for the harsh processing characteristics of modern food manufacturing, including high thermal processing, high-pressure homogenization, emulsification, and prolonged storage, which lead to the degradation of DNA fragments (<200 bp), limiting the utility of conventional detection methods. qPCR assays have consistently achieved limits of detection (LoD) in the low picogram range, even in heavily processed products. For instance, Volkandari et al. (2025) demonstrated sensitive and specific porcine identification across diverse commercial products using ND4L-targeted primers, attaining an LoD of 5 pg/μL with no matrix inhibition in the assay. Similarly, qPCR assays resolved porcine DNA in 131 Korean processed foods with an equivalent sensitivity of 5 pg, maintaining single melt peaks and > 95% amplification efficiency (Kim, Lee, & Lee, 2023). Complementing these findings, our prior investigation employing the mtDNA ND2 gene biomarker successfully amplified bovine and porcine DNA, achieving amplicons from 0.01 pg of template DNA (Barido et al., 2023). The performance of both PCR and qPCR depends critically on primer design. Primers are short single-stranded oligonucleotides that bind selectively to the target DNA sequence, defining the amplified region and determining assay specificity (Bustin & Huggett, 2017; Cahyadi et al., 2019). Mitochondrial DNA is widely preferred as a basis for primer design owing to its high copy number per cell, maternal inheritance, and considerable inter- and intraspecies sequence divergence, which enables reliable species discrimination, even in highly degraded samples (Dong et al., 2021; Kumar et al., 2022; Shoop et al., 2023). Members of this 16 kb genome have been widely used for species identification in various matrices, including ND2 (Barido et al., 2023; Mohamadzade Namin et al., 2022), ND4L (Volkandari et al., 2025), cytochrome b (Ali, Hashim, Mustafa, Man et al., 2012; Novianty et al., 2017), 12S rRNA (Cahyadi et al., 2020; Galal-Khallaf, 2021), and 16S rRNA (Fernandes et al., 2018). Despite the widespread use of mitochondrial genes for species authentication, existing assays are often constrained by large product sizes (>200 bp), susceptibility to non-specific cross-reactivity with phylogenetically proximal taxa, and reduced amplification efficiency in thermally processed food matrices (Nehal et al., 2021). For example, while primers derived from mtDNA cytochrome b and D-loop exhibit commendable sensitivity in fresh templates, they demonstrate substantially attenuated amplification in commercial products subjected to highly processed samples. Therefore, studies have recommended a product size in the range of 50–150 bp to have the highest sensitivity and specificity (Minoudi et al., 2025; Wang et al., 2021; Sarlak et al., 2022). The COX3 encodes a catalytic subunit of cytochrome c oxidase (Complex IV), the terminal enzyme of the mitochondrial electron transport chain (Kadenbach & Hüttemann, 2015). As a protein-coding gene residing on the heavy strand of the mitochondrial genome, COX3 benefits from the inherently high copy number of mtDNA per cell, ranging from hundreds to thousands, depending on the tissue type (Mohamed Yusoff et al., 2025; Shoop et al., 2023). This supports reliable amplification, even when DNA is substantially degraded, as commonly encountered in heat-processed food matrices. Unlike cytochrome b or ND2, which have been more widely adopted in food authentication (Ali, Hashim, Mustafa Man et al., 2012; Ali, Hashim, Mustafa, & Che Man, 2012; Ni'mah et al., 2016; Yacoub et al., 2017), COX3 harbors sufficient interspecies sequence divergence in the primer annealing regions to support species-specific primer design for Gallus gallus domesticus, while showing low intraspecific variability across breeds. To the best of our knowledge, COX3 has not previously been applied for food species authentication, despite its established use in phylogenetic studies (Tian, Liu et al., 2013; Tian, Luo et al., 2013; Yang et al., 2022). The COX3 is generally more divergent than cytochrome c oxidase subunit I (COI) and cytochrome b (Hikosaka et al., 2010). Therefore, this study aimed to design and validate species-specific primers targeting COX3 for the detection of chicken DNA in food products using quantitative PCR (qPCR). Materials and methods Sample preparation Raw meat samples were purchased from licensed commercial supermarkets in Yogyakarta, Indonesia, to obtain commercially available meat products and representative target and non-target species for primer validation. Chicken meat was used as a positive control for COX3-targeted amplification, whereas beef, pork, sheep, duck, shrimp, and squid were used as negative controls to evaluate primer specificity. Approximately 100 mg of muscle tissue was collected from each sample using sterile scalpels. All procedures were performed in triplicate to ensure statistical robustness and analytical reproducibility. Samples were immediately transferred to sterile 2.0 mL microcentrifuge tubes, individually sealed to prevent cross-contamination, and stored at −20 °C until DNA extraction. The relative LoD of the COX3 primers was determined using a binary DNA admixture series (chicken:beef). DNA extracts from pure tissues were mixed in varying proportions to achieve concentrations ranging from 0.1 to 100% (v/v), with a standardized total volume of 500 μL for each admixture (Table 1). These mixtures were used as templates to evaluate the sensitivity of the primers in detecting target DNA within a non-target DNA background. Table 1Preparation of binary DNA admixtures (chicken in beef) for relative LoD determination.Chicken Content in Beef Admixture (v/v)Chicken DNA Volume (μL)Beef DNA Volume (μL)Total volume (μL)0.1%0.5499.55000.5%2.5497.55001%54955005%2547550010%5045050030%15035050050%25025050070%35015050090%45050500100%5000500 Preparation of binary DNA admixtures (chicken in beef) for relative LoD determination. DNA extraction and quantification Total genomic DNA was extracted from both reference and commercial samples using the gSYNC™ DNA Extraction Kit (Geneaid Biotech Ltd., New Taipei City, Taiwan), following the manufacturer's protocol. Briefly, 100 mg of homogenized muscle tissue or processed meat products was subjected to enzymatic lysis using Proteinase K and GS Buffer, followed by incubation at 60 °C until complete tissue lysis was achieved. The lysate was treated with absolute ethanol to precipitate the DNA and subsequently transferred to a gSYNC™ Mini Column. After a series of washing steps using W1 and Wash Buffers to remove contaminants and inhibitors, the purified DNA was eluted in 50 μL of Elution Buffer. The concentration and purity (A260/A280 ratio) of the extracted DNA were determined using a Multiscan G0 Microplate Spectrophotometer (Thermo Fisher Scientific, Vantaa, Finland). All DNA extracts were stored at −20 °C for further use. Specific primer design Primers targeting mtDNA COX3 in chickens (Gallus gallus domesticus; accession no. OP740917.1) were designed using the Primer-BLAST tool of the National Center for Biotechnology Information (NCBI) (https://www.ncbi.nlm.nih.gov/tools/primer-blast/). Primer design parameters were optimized according to the Minimum Information for Publication of Quantitative Real-Time PCR Experiments (MIQE) guidelines (Bustin et al., 2009). These criteria included a primer length of 18–25 nucleotides, GC content of 40–60%, and Tm of approximately 58-62 °C based on nearest-neighbor thermodynamic calculations. The design also ensured minimal secondary structures and primer dimers. Following initial design, the thermodynamic properties of the selected primer pair, including hairpin formation, self-dimer stability, and Gibbs free energy (ΔG), were evaluated using four independent web-based platforms: IDT OligoAnalyzer (Integrated DNA Technologies, Coralville, IA, USA; https://www.idtdna.com/pages/tools/oligoanalyzer), NetPrimer (Premier Biosoft, Palo Alto, CA, USA; http://www.premierbiosoft.com/netprimer/), Multiple Primer Analyzer (Thermo Fisher Scientific, Waltham, MA, USA), and OligoEvaluator™ (Sigma-Aldrich/Merck KGaA, Darmstadt, Germany). Cross-platform validation was conducted to minimize algorithm-dependent bias in thermodynamic estimates, as ΔG values vary substantially across tools owing to differences in parameter sets and calculation conditions (Owczarzy et al., 2008). A complete cross-platform comparison is provided in Supplementary Table S1. The amplicon length was constrained to 70–120 bp to ensure efficient amplification of fragmented mitochondrial DNA from processed meat matrices. The best primers were commercially synthesized by 1st BASE (Singapore), reconstituted in nuclease-free water to prepare stock and working solutions, and subsequently used for qPCR amplification of chicken DNA in raw and processed samples. qPCR amplification targeting COX3 Quantitative real-time PCR (qPCR) for the specific detection of chicken COX3 was performed using SYBR Green I chemistry on a Rotor-Gene Q thermal cycler (Qiagen, Hilden, Germany). All reactions were performed in triplicate (n = 3). Each 20 μL reaction mixture contained 10 μL of 2 × SensiFAST SYBR No-ROX Master Mix (Bioline, Meridian Bioscience, UK), 0.8 μL of forward primer (10 μM), 0.8 μL of reverse primer (10 μM), 1 μL of DNA template (20 ng/μL), and 7.4 μL of nuclease-free water (Qiagen, Germany). The thermal cycling conditions were optimized for the COX3 primer pair as follows: initial polymerase activation at 95 °C for 3 min, followed by 30 cycles of denaturation (95 °C for 5 s), annealing (65 °C for 10 s), and extension (72 °C for 15 s). Product specificity was confirmed through post-amplification melt curve analysis from 72 to 95 °C, with a temperature increment of 0.22 °C per step and a 1 s hold time for fluorescence acquisition on the Green Channel. No-template controls (NTCs), with nuclease-free water replacing the template, were included in each run to monitor potential reagent contamination and primer-dimer formation. Cq and Tm values were determined using the Rotor-Gene QREX software v2.3.1 (Qiagen, Germany). To ensure the absence of non-specific products, amplification results were verified using 3% (w/v) agarose gel electrophoresis (100 V for 45 min). The gels were stained with Diamond Nucleic Acid Dye (Promega Corporation, Madison, WI, USA), visualized by UV transillumination (Gel Doc XR+, Bio-Rad Laboratories, Inc., Hercules, CA, USA). Assay performance validation The performance of the COX3 qPCR assay was validated in strict accordance with the Minimum Information for Publication of Quantitative Real-Time PCR Experiments (MIQE) guidelines (Bustin et al., 2009). The evaluated parameters included specificity, repeatability (precision), amplification efficiency (E), linearity (R2), and LoD. Repeatability and precision Assay precision was determined through triplicate amplifications (n = 3 technical replicates) of 20 ng/μL DNA templates in a single experiment. The relative standard deviation (RSDr) was calculated to assess repeatability using the following equation: RSDr (%) = (SD/mean Cq) x 100. Following Broeders et al. (2014), the acceptance threshold for reproducibility was set at RSDr <25%, although a more stringent internal target of <5% was pursued for high-fidelity detection. Sensitivity, linearity, and efficiency The sensitivity of the COX3 primer set was evaluated by establishing absolute and relative LoDs. The absolute LoD was determined using a six-point, 10-fold serial dilution of Gallus gallus domesticus DNA, spanning a concentration range of 100.900 to 1.009 pg/μL. Concurrently, the relative LoD was established using a binary DNA admixture series (chicken in beef) ranging from 0.1% to 100% (v/v) to simulate economically motivated adulteration (EMA). Standard curves were constructed by plotting Cq values against log10 of the DNA concentration to assess the linear dynamic range. Amplification efficiency (E) was derived from the slope of the regression line using the equation described by Bustin et al. (2009). A robust assay was defined by an E value between 90 and 110% (slope between −3.1 and − 3.6) and a correlation coefficient (R2 ≥ 0.98). Analytical specificity The specificity of the primers was cross-validated against six non-target species: beef (Bos taurus), pork (Sus scrofa domesticus), sheep (Ovis aries), duck (Anas platyrhynchos domesticus), shrimp (Litopenaeus vannamei), and squid (Loligo forbesi). Gallus gallus domesticus DNA was used as a positive control. All tests were performed in triplicate to ensure diagnostic consistency. Application to commercial meat products To evaluate the practical applicability of the developed qPCR assay, 17 commercial meat products were sourced from local supermarkets and traditional markets in Yogyakarta, Indonesia. The samples primarily consisted of processed meat products, such as sausages (9) and chicken floss (8). The procedures for sample preparation, handling, DNA extraction, and qPCR analysis were conducted in strict accordance with the protocols described in the previous sections. To ensure consistency and reliability, each commercial sample was analyzed in triplicate (n = 3). Positive results were identified based on the presence of a characteristic amplification curve and a specific melting peak of chicken in the COX3 primer at Tm = 81.3 ± 0.5 °C. Results and discussion COX3 primer design Species-specific primers targeting mitochondrial COX3 were successfully designed based on the genetic sequence of Gallus gallus domesticus (accession no. OP740917.1) using the NCBI Primer-BLAST tool. This design yielded a forward primer (5’-TCACCGTTACATGGGCTCAC-3′; GC content 55%; Tm 60.2 °C; 20 bp) and a reverse primer (5’-AGAGTTAGTGCGTGGATGGC-3′; GC content 55%; Tm 60.1 °C; 20 bp) that generated a 70 bp amplicon (Table 2, Supplementary Fig. S1). The GC content for both primers was optimized at 55%, falling within the ideal range of 40–60% to provide sufficient thermal stability for robust primer-template hybridization while simultaneously minimizing the risk of secondary structure formation or non-specific binding (Dieffenbach et al., 1993). Furthermore, the calculated Tm were 60.2 °C and 60.1 °C for the forward and reverse primers, respectively. The negligible Tm difference (∆Tm = 0.1 °C) is a critical factor for ensuring the synchronized annealing of both primers during the cycling process, thereby optimizing polymerase efficiency and amplification yield. Table 2Physicochemical properties of the COX3 primer set.PrimerSequence (5′ ➔ 3′)Length (bp)Amplicon size (bp)GC (%)Tm (°C)aHairpin ∆G (kcal/mol)bSelf-dimer ∆G (kcal/mol)bForwardTCACCGTTACATGGGCTCAC20705560.04Not detected−5.38ReverseAGAGTTAGTGCGTGGATGGC205560.11Not detectedNot detectedaMelting temperature calculated by NCBI Primer-BLAST.bGibbs free energy (∆G) calculated by NetPrimer (Premier Biosoft), the acceptable threshold (ΔG > −9 kcal/mol). Physicochemical properties of the COX3 primer set. Melting temperature calculated by NCBI Primer-BLAST. Gibbs free energy (∆G) calculated by NetPrimer (Premier Biosoft), the acceptable threshold (ΔG > −9 kcal/mol). This ultra-short amplicon size was strategically selected to facilitate the detection of highly fragmented DNA typically found in heat-processed meat matrices, while maintaining high amplification efficiency and minimizing the formation of primer dimers. Amplicons within the 50–80 bp range can greatly improve the success rate of identifying and authenticating target species in complex samples, similar to the effectiveness observed in edible oils (Chen et al., 2025). Thermodynamic evaluation using OligoAnalyzer (IDT, Singapore) revealed favorable stability, with hairpin ΔG values ranging from −1.39 to 0.09 kcal/mol. The positive or near-zero ΔG values indicate that these secondary structures are thermodynamically unstable at the optimized reaction temperature (65 °C), and therefore did not competitively interfere with primer-template binding (Owczarzy et al., 2008). This thermodynamic profile ensured that the primers remained in a linear, single-stranded conformation, maximizing their accessibility to the target sequence and maintaining a high amplification efficiency throughout the cycling process (Bustin et al., 2009). The predicted Tm was 78.5 °C, satisfying the MIQE guidelines for high-efficiency primer design. Thermodynamic evaluation of self-dimer formation was conducted using NetPrimer (Premier Biosoft), which reported self-dimer ΔG values of −5.38 kcal/mol for the forward primer and no dimer detected for the reverse primer, both well within the acceptable threshold of ΔG > −9 kcal/mol (Owczarzy et al., 2008). A comparative analysis using four independent platforms (IDT OligoAnalyzer, Multiple Primer Analyzer, PrimerStat, and OligoEvaluator™) is presented in Supplementary Table S1. The inter-platform discrepancies observed are consistent with the known differences in thermodynamic algorithm implementations and do not affect our conclusions. Empirical validation confirmed the absence of meaningful primer-dimer interference under assay conditions: NTC reactions produced flat amplification curves with no signal within 30 cycles, melt curve analysis showed no NTC peak, and gel electrophoresis confirmed a single clean 70 bp band with no sub-50 bp products across all lanes, including NTCs (Fig. 1a–c). Characterization of the COX3 primer set for chicken DNA detection. (a) Representative amplification curve showing the Cq; (b) Melting curve analysis indicating a single peak at Tm 81.27 ± 0.05 °C; (c) Agarose gel electrophoresis (2%) confirming the 70 bp amplicon size of the PCR products. Subsequently, extensive in silico validation using NCBI Primer-BLAST was conducted against the nucleotide collection (nt) database to evaluate the quality and selectivity of the primers (Ye et al., 2012). The evaluation results confirmed the exclusive homology of the primers to Gallus gallus domesticus COX3, with 100% identity and zero mismatches found for the target template (Table 3). Specificity analysis against the non-redundant (nt) database further demonstrated high selectivity, as no target templates were found for non-target species, including Bos taurus, Sus scrofa, and various other animals and plants (Table 4). This molecular barrier effectively prevents cross-reactivity and ensures high assay selectivity, which is crucial for the authentication of chicken-based products in complex food matrices. Table 3Sequence validation and homology of the COX3 amplicon.GenePCR product (5′ ➔ 3′)The number of targets recognizedGallus gallus domesticusNNon-targetmtDNA COX3TCACCGTTACATGGGCTCACCACAGCATCACAGAAGGAAACCGAAAACAAGCCATCCACGCACTAACTCTQuery cover 100%Percent Identify 100%E value 6e-291000 Sequence validation and homology of the COX3 amplicon. Table 4In silico specificity of the COX3 primer set.Common nameSpeciesBLAST Database (nt)Predicted AmplificationAnimalsChickenGallus gallus domesticusYesPositive (Target Found)TurkeyMeleagris gallopavoYesNegative (No Mitochondrial Target Found)aGuinea fowlNumida meleagrisYesPositive (Target Found, Non-target)bCattleBos taurus, Bos indicusYesNegative (No Target Found)YakBos grunniensYesNegative (No Target Found)GoatCapra hircusYesNegative (No Target Found)SheepOvis ariesYesNegative (No Target Found)DuckAnas platyrhynchos f. domesticaYesNegative (No Target Found)Pangasian CatfishPangasius spYesNegative (No Target Found)DogCanis lupus familiarisYesNegative (No Target Found)Salmon fishOncorhynchus nerkaYesNegative (No Target Found)Spanish mackerelScomberomorus commersoniiYesNegative (No Target Found)SquidLoligo spYesNegative (No Target Found)Pasific mackerelScomber japonicusYesNegative (No Target Found)RatRattus norvegicusYesNegative (No Target Found)OctopusOctopus vulgarisYesNegative (No Target Found)CamelCamelus bactrianusYesNegative (No Target Found)CarpCyprinus carpioYesNegative (No Target Found) PlantsPaddyOryza sativaYesNegative (No Target Found)CornZea maysYesNegative (No Target Found)WheatTriticum aestivumYesNegative (No Target Found)TomatoSolanum lycopersicumYesNegative (No Target Found)Chili/Bell pepperCapsicum annumYesNegative (No Target Found)Black pepperPiper nigrumYesNegative (No Target Found)CoconutCocos nuciferaYesNegative (No Target Found)SoyGlycine maxYesNegative (No Target Found)SorghumSorghum bicolorYesNegative (No Target Found)aPredicted amplification products for Meleagris gallopavo (842 bp and 1925 bp) were identified in nuclear mitochondrial DNA segments (NUMTs) integrated into the nuclear genome (chromosome Z and 8), not from authentic mitochondrial COX3 sequences. These products are not expected to amplify under the assay conditions used in this study, given that food-derived DNA is typically fragmented below 200 bp.bA predicted 70 bp amplification product was identified in the mitochondrial genome of Numida meleagris, suggesting potential in silico cross-reactivity. Experimental validation against guinea fowl DNA is recommended in future studies. In silico specificity of the COX3 primer set. Predicted amplification products for Meleagris gallopavo (842 bp and 1925 bp) were identified in nuclear mitochondrial DNA segments (NUMTs) integrated into the nuclear genome (chromosome Z and 8), not from authentic mitochondrial COX3 sequences. These products are not expected to amplify under the assay conditions used in this study, given that food-derived DNA is typically fragmented below 200 bp. A predicted 70 bp amplification product was identified in the mitochondrial genome of Numida meleagris, suggesting potential in silico cross-reactivity. Experimental validation against guinea fowl DNA is recommended in future studies. COX3 primer repeatability The intra-assay repeatability of the COX3 primers was evaluated by amplifying chicken DNA (20 ng/μL) across six technical replicates in a single Rotor-Gene Q run (Salamah et al., 2019; Volkandari et al., 2025). Precision was assessed based on the consistency of the Cq and Tm values according to the MIQE guidelines (Bustin et al., 2009). The assay demonstrated excellent primer performance, yielding an average Cq of 13.19 ± 0.20 and a Tm of 81.27 ± 0.05 °C (Table 5, Fig. 1a–c). Table 5Cq values and Tm (°C) for COX3 primers.ReplicationCqTm112.9681.2213.2381.3312.9681.3413.3581.3513.4481.2613.1981.3Average13.1981.27Standard deviation0.200.05RSDr (%)1.500.06RSDr is relative standard deviation of repeatability. Cq values and Tm (°C) for COX3 primers. RSDr is relative standard deviation of repeatability. The RSDr was 1.50% for Cq and 0.06% for Tm. These values were significantly lower than the 25% RSDr acceptance threshold defined by ISO 20813:2019 and MIQE for qPCR, confirming the high validity of this method (Bustin et al., 2009; Broeders et al., 2014; Taylor et al., 2010). The results of this study showed lower variability than those of a previous study utilizing ND4L with RSDr Cq 2.41%, Tm 0.07% (Volkandari et al., 2025). This confirms that the COX3 primer set is a highly reproducible tool for the authentication of chicken DNA. The melt curve profiles (Fig. 1b, 2b, 3b, Fig. 4, Fig. 5) revealed minor peaks in the 74–80 °C region preceding the dominant amplicon peak. These minor peaks are attributable to differential domain melting arising from pronounced intra-amplicon GC heterogeneity and should not be interpreted as evidence of non-specific amplification or primer-dimer formation. Sequence analysis of the 70 bp COX3 amplicon, conducted by dividing the amplicon into consecutive 10-nucleotide sub-regions following Abtahi et al. (2011), revealed substantial variation in GC content across sub-domains: positions 11–20 (GC = 60%) and 51–60 (GC = 70%) represented the highest-GC sub-regions, flanking a markedly low-GC central zone at positions 43–50 (mean GC ≈ 15%), with the overall intra-amplicon GC range spanning 40 percentage points across sub-regions (Supplementary Table S2). This pronounced GC heterogeneity creates discrete melting domains that dissociate sequentially during the temperature ramp, producing minor pre-peak fluorescence transitions before the dominant 81.27 °C peak, consistent with the bimodal melting phenomenon described for short amplicons with asymmetric GC distribution (Abtahi et al., 2011). The fluorescence transitions associated with these subdomain melting events are further explained by the non-saturating intercalation behavior of SYBR Green I, whereby dye molecules released from lower-melting AT-rich regions redistribute to the remaining double-stranded segments, transiently modulating the fluorescence signal during the melting ramp (Zipper et al., 2004). The consistent presence of these transitions across all replicates and their complete absence in NTC reactions confirmed that they reflected the intrinsic melting thermodynamics of the COX3 amplicon sequence, not non-specific amplification products. Analytical specificity of the COX3 primer set. (a) Amplification curves showing positive detection for chicken DNA with no cross-amplification observed in non-target species (beef, pork, sheep, squid, shrimp, and duck); (b) Corresponding melting curve analysis confirming the specificity of the amplicon for Gallus gallus domesticus. All samples were analyzed in triplicate. Sensitivity analysis of the COX3 qPCR assay (Absolute LoD). (a) Amplification curves generated from 10-fold serial dilutions of chicken DNA. The absolute functional LoD was established at 10.09 pg/μL. (b) Melting curve of PCR product. (c) Standard curve showing high linearity (R2 = 0.990) and 89% PCR efficiency. All samples were analyzed in triplicate. Sensitivity analysis of the COX3 assay in binary DNA mixtures. (a) Amplification curves and (b) Melting curves of DNA mixtures with simulated contamination ranging from 100% to 0.5% (v/v) chicken DNA in a beef DNA background. (c) Standard curve showing high linearity (R2 = 0.999) and 100% PCR efficiency. All samples were analyzed in triplicate. Detection of chicken DNA in commercial processed products. (a and c) Representative qPCR amplification curves of sausage and meat floss products, respectively and (b-d) derivative melt peak analysis showing the identification of undeclared chicken DNA in beef-labeled sausages and meat floss. Positive signals in samples declared as 100% beef indicate a high incidence of mislabeling (41.2% of total samples; 63.6% of beef-labeled products). NTC: Non-Template Control. All samples were analyzed in triplicate. COX3 primer specificity Primer specificity is defined as the accurate and exclusive amplification of target DNA while failing to amplify non-target templates (Premanandh, 2013). This parameter is a primary determinant of qPCR analytical fidelity, especially in heterogeneous food matrices, where cross-reactivity can invalidate species authentication results. In this study, the COX3 primers were rigorously validated against six non-target species, including beef (Bos taurus), pork (Sus scrofa domesticus), sheep (Ovis aries), duck (Anas platyrhynchos domesticus), shrimp (Litopenaeus vannamei), squid (Loligo forbesi), and chicken (Gallus gallus domesticus), as positive controls. The assay demonstrated no amplification of any of the six non-target species within 30 cycles of the qPCR run (Table 6, Fig. 2a, b). Conversely, chicken DNA yielded robust and consistent amplification with a mean Cq of 14.46 ± 0.08 and a distinct Tm of 81.55 ± 0.21 °C. The low standard deviation of both Cq and Tm values indicated high technical precision and target stability, meeting the stringent requirements of MIQE guidelines (Bustin et al., 2009). Table 6Analytical specificity of the COX3 primer set against target and non-target species.SpeciesCommon NameMean Cq ± SDMean Tm ± SD (°C)ResultsGallus gallus domesticusChicken14.46 ± 0.0881.55 ± 0.21(+)Bos taurusBeefN·D.N·D.(−)Sus scrofa domesticusPigN.D.N.D.(−)Ovis ariesSheepN.D.N.D.(−)Anas platyrhynchosDuckN.D.N.D.(−)Litopenaeus vannameiShrimpN.D.N.D.(−)Loligo spSquidN.D.N.D.(−)All analyses were performed in triplicate technical replicates.(+) is positive result; chicken DNA detected (amplification curve observed with a specific melting peak at Tm = 81.55 ± 0.21 °C within 30 PCR cycles)(−) is negative result; chicken DNA not detected (no amplification observed within 30 PCR cycles and no specific derivative melting peak detected)N.D. is not detected. Analytical specificity of the COX3 primer set against target and non-target species. All analyses were performed in triplicate technical replicates. (+) is positive result; chicken DNA detected (amplification curve observed with a specific melting peak at Tm = 81.55 ± 0.21 °C within 30 PCR cycles) (−) is negative result; chicken DNA not detected (no amplification observed within 30 PCR cycles and no specific derivative melting peak detected) N.D. is not detected. These experimental findings were further corroborated by a comprehensive in silico analysis (Table 4), which predicted negative amplification in 25 animal and plant species. The absence of cross-reactivity with ducks (Anas platyrhynchos domesticus) is particularly significant, despite both species belonging to the Galloanserae superorder (Kimball et al., 2019). However, expanded in silico analysis of additional avian species revealed findings that warrant further discussion. In silico analysis identified predicted amplification products of 842 and 1925 bp in the nuclear genome of Meleagris gallopavo, likely representing NUMTs rather than authentic mitochondrial targets. As these product sizes far exceed both the 70 bp mitochondrial amplicon and the typical DNA fragment size in processed food matrices (<200 bp), cross-reactivity with turkey DNA under the assay conditions used is considered unlikely. Experimental validation against turkey DNA is recommended in future studies, particularly given the commercial prevalence of turkey-containing products in international markets. In silico analysis also revealed that the COX3 primer set may amplify a 70 bp product from Numida meleagris (guinea fowl) mitochondrial DNA, with three mismatches in the forward and 2–3 mismatches in the reverse primer annealing regions. Although guinea fowl (Numida meleagris) is commercially utilized as a niche poultry species in certain African and European markets (Bucław et al., 2024), it is not commonly encountered in the Indonesian processed meat market, where poultry production and consumption are predominantly dominated by chickens and ducks. In Indonesia, guinea fowl are more frequently maintained in small-scale or ornamental poultry systems rather than in intensive commercial meat production (Rahman et al., 2025). Therefore, the practical relevance of this potential cross-reactivity is expected to be limited to the intended application context of the present study. This specificity is attributed to the strategic positioning of the ‘3’-termini of the primers at interspecies single nucleotide polymorphisms (SNPs), which impose a thermodynamic penalty that prevents non-target elongation (Stadhouders et al., 2010). The zero-mismatch strategy for the target (chicken) and deliberate mismatch positioning for non-targets ensure that this assay remains robust even under suboptimal annealing temperatures, which are often encountered in routine food safety laboratories. The lack of signals for common meat adulterants (beef, pork, and sheep) and seafood components (shrimp and squid) validates this method as a reliable tool for food authentication in complex multi-ingredient products. In comparison to multiplex PCR approaches, which simultaneously target multiple species using various mitochondrial markers. Cheng et al. (2022) developed a multiplex PCR with target species such as camel, cattle, dog, pig, chicken, sheep and duck for the simultaneous identification of seven species. In addition, Yang et al. (2022) constructed a multiplex PCR that could authenticate eight species: ostrich (753 bp), cat (564 bp), goose (391 bp), duck (347 bp), chicken (268 bp), horse (227 bp), dog (190 bp), and sheep (131 bp). More recently, Villanueva-Zayas et al. (2026) developed a methodologically distinct but complementary COI metabarcoding approach (COI-MBC) employing a universal 201 bp primer pair for simultaneous multi-species authentication of meat products, demonstrating the successful identification of major meat species. However, the authors reported a quantitative bias in pork-turkey mixtures and recommended the use of taxon-specific dual primer sets (targeting mammals and poultry separately) to improve quantitative reliability, highlighting an inherent limitation of universal primer-based approaches in heterogeneous matrices. The COX3-based single-species qPCR assay offers greater specificity, making it particularly suitable for targeted species verification and regulatory enforcement in the Indonesian food authentication context. Sensitivity (LoD) The LoD is defined as the lowest DNA concentration that can be reliably detected, which is essential for identifying species in heat-treated meat products, where DNA is typically fragmented. In this study, the sensitivity of the COX3 primer set was validated using both absolute DNA concentrations and relative percentages in binary meat mixtures. This dual approach was employed to assess assay performance across varying sample purities and simulate real-world adulteration scenarios during food inspections (Rohman et al., 2011). The absolute LoD was evaluated through a 10-fold serial dilution of Gallus gallus domesticus DNA, ranging from 100.900 ng/μL to 1.009 pg/μL (Table 7, Fig. 3a and b). The assay exhibited high linearity, with a correlation coefficient (R2) of 0.998 and PCR efficiency of 89% (Fig. 3c). Reliable amplification was achieved down to 1.009 pg/μL with a mean Cq of 29.06 ± 0.65. While the Tm was stable at higher DNA concentrations (81.2 °C), the derivative melt peak was not resolved at the 1.009 pg/μL threshold. Therefore, following a rigorous validation approach, the functional absolute LoD was established as 10.09 pg/μL for this assay. At this level, the assay provided consistent amplification (Cq 23.86 ± 0.42) and a clearly resolved, specific Tm peak (81.55 ± 0.07 °C). Consequently, this concentration was defined as the functional LoD for the assay. This functional LoD (10.09 pg/μL) demonstrates superior sensitivity compared to the COI-based assay reported by Kusnadi and Harfiyanti (2023), which showed an LoD of 50 ng/μL. In addition, a recent study by Villanueva-Zayas et al. (2026) developed a COI metabarcoding approach (COI-MBC) and reported a relative detection threshold of 1% (v/v). Under the experimental conditions evaluated in the present study, the COX3-based qPCR assay achieved a functional relative LoD of 0.5% (v/v), accompanied by high quantitative precision (CV < 6%). Although direct comparisons between metabarcoding and qPCR platforms should be interpreted with caution due to methodological differences, these results indicate that the COX3-based assay is capable of detecting target material at lower relative proportions than those reported for the COI-MBC approach. Table 7Absolute LoD of the COX3 primer set.DNA concentration (pg/μL)Average Cq ± SDCV (%)Average Tm ± SDCV (%)100.9009.42 ± 0.353.6881.20 ± 0.100.1210.09012.13 ± 0.625.1281.27 ± 0.120.14100915.71 ± 0.372.3881.27 ± 0.120.14100.920.54 ± 0.472.2981.30 ± 0.170.2110.0923.86 ± 0.421.7481.55 ± 0.070.091.00929.06 ± 0.652.23N.D.N.D.N.D. is not detected (no amplification observed within 30 PCR cycles and no resolved derivative melting peak detected). Absolute LoD of the COX3 primer set. N.D. is not detected (no amplification observed within 30 PCR cycles and no resolved derivative melting peak detected). Furthermore, the sensitivity of the COX3 primer set is highly competitive with advanced detection methods; it matches the LoD of 10 pg/μL achieved using TaqMan real-time PCR targeting the Actb gene (Xiang et al., 2017) and is comparable to the 10.24 pg LoD reported for the cytochrome b gene using SYTO-9 staining (Dalsecco et al., 2018). Zhang et al. (2020) employed high-throughput sequencing combined with a custom mitochondrial genome database containing 2354 complete sequences for simultaneous identification of multiple animal species in meat products. While this approach offers broad taxonomic coverage, it requires specialized sequencing infrastructure and bioinformatic pipelines that are not routinely available in Indonesian food-safety laboratories. In contrast, the COX3-based qPCR assay is compatible with standard real-time PCR equipment and provides quantitative data with a lower LoD of 10.09 pg/μL. To simulate real-world food adulteration, a relative LoD was established using binary mixtures of chicken DNA in a beef DNA background ranging from 0.1% to 100% (v/v) (Table 8, Fig. 4a and b). The assay demonstrated exceptional performance in this complex matrix, achieving 100% PCR efficiency and an R2 value of 0.999 (Fig. 4c). These kinetics demonstrate that the high concentration of background beef DNA does not inhibit Taq polymerase activity, ensuring that the COX3 primers maintain optimal binding and extension throughout the cycling process. Table 8Relative LoD of the COX3 primer set.Adulterated DNA (%)Average Cq ± SDCV (%)Average Tm ± SDCV (%)0.1%26.12 ± 0.391.49N.D.N.D.0.5%23.55 ± 0.190.7982.23 ± 0.150.191%22.68 ± 0.050.2382.23 ± 0.150.195%20.24 ± 0.130.6581.9 ± 0.260.3210%19.50 ± 0.110.5581.80 ± 0.170.2130%17.86 ± 0.170.9781.67 ± 0.210.2550%16.95 ± 0.140.8481.60 ± 0.170.2170%16.84 ± 0.080.4581.63 ± 0.150.1990%16.33 ± 0.080.4981.67 ± 0.310.37100%16.38 ± 0.392.3781.57 ± 0.120.14N.D. is not detected (no amplification observed within 30 PCR cycles and no resolved derivative melting peak detected). Relative LoD of the COX3 primer set. N.D. is not detected (no amplification observed within 30 PCR cycles and no resolved derivative melting peak detected). The assay demonstrated high linearity and precision across the dilution series. At the lowest concentration tested (0.1%), a Cq was observed at 26.12 ± 0.39, with a low coefficient of variation (CV) of 1.49%. However, a stable and resolvable derivative melting peak (Tm) was not observed at this concentration. In contrast, at 0.5% (v/v), the assay yielded both a robust Cq signal (23.55 ± 0.19) and a clearly defined Tm peak of 82.23 ± 0.15 °C. Based on these results, the functional relative LoD of this assay was defined as 0.5% (v/v). This threshold ensures the detection of target DNA and its definitive species authentication through melt curve analysis. This sensitivity level (0.5%) is highly competitive for food authentication. Furthermore, our COX3-based assay exceeded the 1% LoD reported by Dalsecco et al. (2018) using the cytochrome b gene and SYTO-9 chemistry. The ability to achieve ddPCR-level sensitivity using a more cost-effective qPCR platform while matching or exceeding the mitochondrial mini-barcodes previously reported in processed meat (Barido et al., 2023; Chen et al., 2025) underscores the robustness of the COX3 marker for routine food safety and halal verification. Application of COX3 primer on commercial products Seventeen commercial meat products from Indonesian markets were screened for chicken DNA using the COX3 quantitative PCR (qPCR) assay. Chicken DNA was confirmed in 13 of 17 samples (76.5%), with Cq values ranging from 9.99 ± 0.53 to 19.43 ± 0.17, and CV < 6%. The stability of the melting temperatures (Tm = 81.40–81.73 °C; CV < 0.5%) confirmed the high reproducibility of the amplicon across diverse food matrices (Table 9, Fig. 5a-d). Table 9Application of COX3 primers for chicken DNA detection in commercial meat products.Products typeDeclared LabelDNA concentration (ng/μL)Mean Cq ± SD⁎CV (%)Mean Tm ± SD⁎CV (%)Label Conformity⁎Sausage1Chicken1021.00 ± 9.2410.86 ± 0.504.5681.43 ± 0.210.26CompliantSausage2Beef, chicken224.70 ± 1.1510.26 ± 0.464.4581.40 ± 0.170.21CompliantSausage3Beef, chicken88.83 ± 3.0310.09 ± 0.232.2581.43 ± 0.210.26CompliantSausage4Beef251.50 ± 1.4410.44 ± 0.514.8981.53 ± 0.210.26Non-compliantSausage5Beef81.12 ± 6.8628.23 ± 0.321.15N.D.–CompliantSausage6Chicken36.89 ± 0.7211.97 ± 0.352.9681.40 ± 0.100.12CompliantSausage7Beef, chicken449.70 ± 1.809.99 ± 0.535.3581.43 ± 0.120.14CompliantSausage8Beef198.40 ± 1.9510.45 ± 0.313.0081.57 ± 0.060.07Non-compliantSausage9Beef, chicken164.40 ± 6.4210.95 ± 0.353.1881.50 ± 0.100.12CompliantFloss1Beef9.75 ± 0.0019.43 ± 0.170.8681.73 ± 0.150.19Non-compliantFloss2Beef34.08 ± 0.5114.71 ± 0.231.5381.60 ± 0.100.12Non-compliantFloss3Beef34.23 ± 0.6515.54 ± 0.080.5481.73 ± 0.060.07Non-compliantFloss4Beef47.14 ± 2.1715.09 ± 0.040.2581.63 ± 0.060.07Non-compliantFloss5Beef77.60 ± 7.4313.49 ± 0.090.6381.73 ± 0.060.07Non-compliantFloss6Beef66.43 ± 1.44N.D.–N.D.–CompliantFloss7Beef39.64 ± 5.99N.D.–N.D.–CompliantFloss8Beef36.12 ± 0.1428.07 ± 0.812.90N.D.–CompliantN.D. is not detected (no amplification observed within 30 PCR cycles and no resolved derivative melting peak detected).SD isstandard deviationCV is coefficient of variation.⁎All analyses were performed in triplicate using 20 ng/μL of total DNA template, values are the mean of triplicate measurements. Application of COX3 primers for chicken DNA detection in commercial meat products. N.D. is not detected (no amplification observed within 30 PCR cycles and no resolved derivative melting peak detected). SD isstandard deviation CV is coefficient of variation. All analyses were performed in triplicate using 20 ng/μL of total DNA template, values are the mean of triplicate measurements. The detection of chicken DNA in a high percentage of samples, particularly those labeled as 100% beef (63.6%; 7 out of 11 samples), underscores the significant incidence of mislabeling in the local market. The relatively low Cq values (all <20) observed across the positive samples suggest substantial inclusion of poultry rather than trace-level cross-contamination from the shared production lines. Minoudi et al. (2025) validated quantitative real-time PCR methods for species authentication in Greek commercial meat products, reporting a deliberate adulteration rate of 16%, with chicken being the most frequently undeclared species. These findings are directly parallel to our results for the Indonesian market. Their work confirmed that targeted single-species qPCR assays remain the most practical and sensitive approach for routine regulatory enforcement, particularly for chicken adulteration surveillance. These results demonstrate that the COX3 marker is a robust and reliable tool for species authentication, capable of overcoming matrix-induced inhibition and DNA degradation in processed commercial food products. Conclusion This study successfully validated a novel mitochondrial COX3-targeted qPCR assay as a robust and sensitive tool for the authentication of the chicken (Gallus gallus domesticus) species. By targeting a short amplicon of 70 bp, the assay demonstrated exceptional resilience against DNA degradation, which is a critical requirement for analyzing thermally processed food. The assay achieved high analytical precision (RSDr <1.5%) and a functional absolute LoD of 10.09 pg/μL, with a stable melting temperature (Tm = 81.55 ± 0.07 °C). Furthermore, the primer design facilitated the detection of chicken DNA at concentrations as low as 0.5% (v/v) in complex binary mixtures with 100% amplification efficiency, successfully distinguishing the target from common meat adulterants and closely related avian species. The ability to match the sensitivity of more complex technologies, such as ddPCR and probe-based systems, while utilizing a more accessible dye-based qPCR platform, positions the COX3 marker as a cost-effective and superior molecular biomarker. These findings provide an important framework for enhancing food adulteration surveillance and regulatory quality assurance protocols in the global meat supply chain industry. Nevertheless, this study has several limitations. The relative LoD of 0.5% (v/v) was determined using binary mixtures of purified DNA extracts rather than model food systems with physically defined meat substitution percentages. The actual detection threshold in processed meat products may differ owing to species-specific DNA extraction efficiency, matrix-induced PCR inhibition, and differential DNA fragmentation during thermal processing. Future studies should validate the COX3 assay performance using laboratory-prepared model meat systems to establish a quantitative w/w substitution threshold that is directly applicable to regulatory enforcement. Declaration of generative AI and AI-assisted technologies in the manuscript preparation process During the preparation of this work, the authors used Paperpal and Claude in order to enhance the flow of the English and ensure the text is free of errors in grammar, spelling, punctuation, and tone. After using this tool, the authors reviewed and edited the content as needed and take full responsibility for the content of the published article. CRediT authorship contribution statement Saradiva Bilqisyah Azka: Writing – original draft, Investigation, Formal analysis, Data curation. Farouq Heidar Barido: Writing – original draft, Validation, Investigation, Formal analysis. Slamet Diah Volkandari: Writing – review & editing, Validation, Supervision, Methodology, Conceptualization. Dicky Tri Utama: Methodology, Formal analysis. Yuny Erwanto: Methodology, Formal analysis. Muhammad Cahyadi: Writing – review & editing, Validation, Supervision, Methodology, Funding acquisition, Conceptualization. Funding This study was fully supported by the RKAT PTNBH fiscal year 2024 through the applied seed research scheme (PUT-UNS) with assignment of agreement decree number 194.2/UN27.22/PT.01.03/2024, Institute for Research and Community Service, Universitas Sebelas Maret, Indonesia. Declaration of competing interest The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper. Acknowledgments The authors would like to convey their gratitude to the Institute for Research and Community Service, Universitas Sebelas Maret, for providing research funding. The authors also gratefully acknowledge the National Research and Innovation Agency of Indonesia (BRIN) for providing the laboratory facilities that supported the experimental work of this study. Footnotes Supplementary data to this article can be found online at https://doi.org/10.1016/j.fochms.2026.100433. Contributor Information Slamet Diah Volkandari, Email: slam022@brin.go.id. Muhammad Cahyadi, Email: mcahyadi@staff.uns.ac.id. Supplementary materials Supplementary material Data availability Data will be made available on request. References Associated Data Supplementary Materials Supplementary material Data Availability Statement Data will be made available on request.