Madura cattle have been recognized by the Indonesian Ministry of Agriculture as one of Indonesia’s local breeds, originating from Madura Island and its surrounding areas. Madura cattle are among Indonesia’s valuable local livestock genetic resources that require protection and conservation (National Standards Agency of Indonesia, 2020). The conservation of Indonesian local cattle is important because these breeds possess unique genetic backgrounds, adaptive traits, and potential breeding value. Recent genetic and genomic studies have emphasized that genetic information is essential for supporting sustainable utilization, conservation, and breeding strategies for the Indonesian local cattle population (Sutarno & Setyawan, 2015). The Madura cattle population on Madura Island makes a significant contribution, meeting approximately 24% of the demand for beef cattle in East Java (Nurgiartiningsih et al., 2016). Additionally, Madura cattle play a vital role as a symbol of identity and culture for the Madura community (Nurgiartiningsih et al., 2016).

The population of crossbred Madura cattle on Madura Island over the past few decades has reached 53.3% of the total cattle population on the island (Nurgiartiningsih et al., 2016). This poses a potential threat to the genetic diversity of local livestock. Uncontrolled crossbreeding is also believed to be a major threat to the sustainability of Indonesia’s local livestock. Uncontrolled crossbreeding with exotic breeds has also been reported as a factor that may contribute to genetic erosion and reduced quality in Indonesian local cattle populations (Sutarno & Setyawan, 2015).

To prevent the loss of local genetic resources, comprehensive genetic information on local cattle breeds is needed as baseline data for developing conservation programs and breeding strategies (Petretto et al., 2022); (Pramod et al., 2018); (Xia et al., 2019). Mitochondrial DNA analysis is frequently used to assess intra- and interspecies diversity, trace maternal lineages, and examine evolutionary relationships among various species (Anwar et al., 2025). Popular mtDNA markers have been used in analyses of Bali cattle breeds (Jakaria et al., 2019), Thailand dairy cattle(Ariyaraphong et al., 2021), Alpine cattle (Granado et al., 2021), and Podolian cattle (Lorenzo et al., 2018), albeit only partially. Currently, intensive research on mitochondrial genomes is being conducted to examine genetic diversity in Uruguayan cattle (Liu et al., 2020), Georgian mountain cattle (Kunelauri et al., 2022), Shigaste humped cattle (Guo et al., 2018), Yunling cattle (Xia et al., 2019), Sardinian cattle(Petretto et al., 2022), Native Korean cattle (Wang et al., 2016), Guangxi cattle(Xia et al., 2019), and Bali cattle (Anwar et al., 2025); (Jakaria et al., 2025). Mitochondrial analysis of Madura cattle has been conducted, but only using partial markers such as D-loop (Hartatik et al., 2015), cytB (Prihandini et al., 2020), and COI (Utami et al., 2024). Although these markers provide useful preliminary information, they are insufficient to describe complete mitochondrial genome organization, nucleotide composition, codon usage, tRNA structure, control region variation, and whole mitogenome-based phylogenetic relationships. No studies have been reported on the analysis of the complete mtDNA genome of Madura cattle. Therefore, this study addresses this gap by providing the first complete mitochondrial genome characterization of Madura cattle using long-read Oxford Nanopore Technology. This study aims to fully characterize the mitochondrial genome of Madura cattle, reveal their genetic diversity, and examine their phylogenetic relationships with other cattle breeds worldwide.

The mitochondrial genome features of Madura cattle, including sequence length, gene organization, and strand distribution, are consistent with those reported in Bos taurus(Estrada et al., 2024), B. indicus(Arya et al., 2025), B. javanicus(Jakaria et al., 2025), Bos grunniens(Gu et al., 2007), Bos gaurus(Kamalakkannan et al., 2020), and Bison bison(Douglas et al., 2011). All of which share a conserved mitochondrial genome length of approximately 16.3 kb, indicating a stable genomic architecture across vertebrates. Protein-coding genes (PCGs) in the mtDNA of Madura cattle show variation in the use of start and stop codons. The start codons in Madura cattle show no differences compared to those of B. javanicus(Anwar et al., 2025), Bos taurus (Arbizu et al., 2022), and B. grunniens (Gu et al., 2007). Meanwhile, the stop codons of the Madura cattle tend to be more varied compared to their codons, namely the stop codon AGA (cytB), TAA (ND6, ND5, ND4L, ATP6, ATP8, COX2, and COX1), TA- (ND1), and T- (ND4, ND3, COX3, and ND2).

The majority of vertebrates use TAA as the primary stop codon in their mitochondrial DNA (mtDNA) genomes. In the standard genetic code, the AGA and AGG codons code for the amino acid arginine, but in vertebrate mtDNA, they function as stop codons(Zhang et al., 2022). The codon AGA also functions as an alternative codon that acts as a translation terminal signal, particularly in the cytB gene, and has been reported in several vertebrates, such as B. javanicus(Anwar et al., 2025), Mustela lutreola(Skorupski, 2022), B. taurus(Estrada et al., 2024), Mesechinus miodon (Kong et al., 2016), and Myotis muricola(Yoon & Park, 2015). The pattern of incomplete stop codon usage (TA- and T--) in Madura cattle is also commonly found in other vertebrates; these codons become complete termination codons (TAA) after undergoing post-transcriptional polyadenylation at the 3’ end of the mRNA (Skorupski, 2022), which is the addition of a poly (A) tail to the mRNA, thereby forming a complete stop codon sequence (Ojala et al., 1981).

No differences were found in the use of stop codons in the mtDNA genome of Madura cattle compared to B. javanicus (Jakaria et al., 2025). However, differences were found in the stop codons of the genes ND3, COX3, and ND2 in B. taurus, especially TA- (Estrada et al., 2024), whereas the stop codons in Madura cattle for those genes are T--. The Madura cattle stop codon differs from that of Bos grunniens, specifically the AGA stop codon, which is absent and replaced by TAG (Gu et al., 2007).

The complete mitochondrial genomes of Madura cattle showed gene content and organization generally comparable to those of other Bovidae species. The difference in sequence length is due to the presence of a tandem repeat motif in the control region (D-loop) with the motif 5’-ATA AGT ACA TAA TAT TTA TGT A-3’ in samples Ma2 and Ma3, which was not found in samples Ma1, Ma4, and Ma5. The repeat motif in Ma2 was duplicated 6 times, whereas in Ma3 it was duplicated only 3 times. The same repeat motif was also found in B. javanicus (Anwar et al., 2025); (Jakaria et al., 2025). This motif was found only in the species B. javanicus. Therefore, it has great potential to serve as a marker for B. javanicus(Jakaria et al., 2025). To date, the function of repetitive motifs in the D-loop region remains poorly understood. However, it is hypothesized that these motifs act as an independent regulatory region that may enhance the efficiency of mtDNA replication and transcription, providing an adaptive advantage through greater flexibility in regulating energy production and cellular metabolism (Kamalakkannan et al., 2020).

The nucleotide composition of the PCGs shows that 60.08% consists of adenine (A) and Thymine/uracil (T/U). Translation results indicate that the amino acid sequence of Madura cattle is generally similar to that of B. indicus, except for the ND3 gene, which is identical to that of B. javanicus. Two amino acid mutation sites were found to be unique to Madura cattle, consisting of a change in the ATP8 and ND5 genes from the amino acid phenylalanine (F) to leucine (L) at position 24, and a mutation from the amino acid valine (V) to isoleucine (I) at position 172.

The AT-skew values tend to be positive, and the GC-skew values tend to be negative, indicating a pattern of base composition asymmetry that reflects a nucleotide composition bias toward the bases adenine(A) and cytosine (C). This is consistent with the general characteristics of mtDNA genomes in mammalian groups, including the Bovidae family (Priyono et al., 2020). At the gene level, AT-skew values vary among genes. In Madura cattle, the COX1, COX3, and ND6 genes exhibit negative AT-skew values, whereas the genes encoding the proteins (ND1, ND2, COX2, ATP8, ATP6, ND3, ND4L, ND4, ND5, and cytB) have positive values. The same pattern has been reported in B. javanicus (Anwar et al., 2025), B. indicus(Arya et al., 2025), Moschiola indica(Sarvani et al., 2018), and Bubalus bubalis(Kumar De et al., 2019). However, in Bubalus quarlesi, the AT-skew pattern exhibits a distinct characteristic; a negative AT-skew value was detected only in the ND6 gene, while other protein-coding genes maintained a positive AT-skew (Priyono et al., 2020). These pattern differences indicate lineage-specific variations in mitochondrial nucleotide composition bias among Bovidae, which likely reflect differences in historical mutation pressure, mtDNA replication dynamics, or distinct maternal evolutionary backgrounds between the genera (Kamalakkannan et al., 2020).

The RSCU pattern in Madura cattle indicates unequal usage of synonymous codons in mitochondrial protein-coding genes. Several codons with relatively high RSCU values ended with adenine, including CGA, CTA, and GTA, suggesting a tendency toward A-ending codon preference. A similar pattern was reported in Bali cattle (B. javanicus), where the codons with the highest RSCU values were CUA/CTA (Leu, 1.92), GUA/GTA (Val, 1.75), GGA (Gly, 1.64), and GCA (Ala, 1.54), and codons with adenine at the third codon position were preferentially used (Jakaria et al., 2025). Therefore, the RSCU pattern of Madura cattle is generally consistent with the A-ending codon preference reported in Bali cattle, although the dominant codons were not identical. The frequent representation of leucine and serine in Madura cattle also agrees with the general amino acid usage pattern commonly observed in bovine mitochondrial protein-coding genes. In Madura cattle, leucine and isoleucine were the two most frequent amino acids, with 532 and 481 codons, respectively. (Arya et al., 2025) also analyzed RSCU patterns and selection pressure in the complete mitochondrial genome of Badri cattle (B. indicus) and reported that mitochondrial protein-coding genes were mainly shaped by purifying selection. This suggests that codon usage variation in bovine mitogenomes may be influenced not only by nucleotide composition bias, but also by functional constraints related to mitochondrial protein translation (Parvathy et al., 2022).

A total of 22 tRNA genes were identified with a total length of 1515 bp. The tRNA anticodons found correspond to those commonly reported in vertebrates(Jakaria et al., 2025). Based on tRNA structure visualization, the Madura cattle tRNA forms a cloverleaf secondary structure consisting of an acceptor stem, D-arm, anticodon arm, TΨC arm, and variable loop. Several mitochondrial tRNAs show deviations from this structure, namely trnS1 (tRNA-Ser1) and trnK (tRNA-Lys) (Figure 4). The structure of trnS1 and trnK differs from other tRNAs due to the loss or instability of the loop on the dihydrouridine (DHU) arm (Watanabe et al., 2014). These findings are consistent with the characteristics of tRNA in vertebrates; the cloverleaf structure and deviations in trnS1 and trnK were also found in Bos gaurus(Kamalakkannan et al., 2020), B. javanicus (Jakaria et al., 2025), and B. taurus(Estrada et al., 2024).

Phylogenetic reconstruction and genetic distance results using the full mitochondrial genome showed that some Madura cattle samples clustered with B. indicus, while others clustered with B. javanicus. This topology indicates the presence of distinct mitochondrial lineages within the analyzed Madura cattle samples, reflecting a complex maternal background related to both B. indicus and B. javanicus lineages. This clustering pattern is also consistent with the theory of mitochondrial genome inheritance, which is passed down exclusively through the maternal line (Ladoukakis & Zouros, 2017). The B. indicus-related lineage may be associated with the historical dissemination of indicine cattle across Asia through trade, migration, and livestock management practices (Pitt et al., 2019), whereas the B. javanicus-related lineage may reflect maternal introgression from banteng-derived cattle populations (Wang et al., 2026). Because mitochondrial DNA is maternally inherited, these findings describe the maternal history of Madura cattle and should be complemented by nuclear genomic data to resolve genome-wide ancestry and admixture patterns further.

The complete mitochondrial genome of Madura cattle was successfully characterized using Oxford Nanopore Technology, revealing the typical bovine mitochondrial organization and mitochondrial variation in the D-loop and protein-coding regions. Genetic distance and phylogenetic analysis consistently identified two maternal lineages, with one closely related to Bos javanicus and the other to Bos indicus. This finding provides baseline information on the maternal genetic diversity of Madura cattle and may support future genetic characterization and conservation. Further studies using broader sampling and nuclear genomic data are needed.

J. Jakaria, an associate editor at the Tropical Animal Science Journal, was not involved in the decision-making process for this publication. The authors also declare no conflicts of interest related to this study.

The authors thank the Ministry of Higher Education, Science, and Technology for funding their research through the Regular Fundamental Research (PFR) scheme, under contract number 22048/IT3.D10/PT.01.03/P/B/2024, and the Master’s Thesis Research Grant (PTM) under the BIMA system, under contract number 18939/IT3.D10/PT.01.03/P/B/2026.

While preparing this work, the authors utilized Grammarly to enhance the writing structure and grammar. Following this, we reviewed and edited the content as necessary and assumed full responsibility for the published material.