The research team, led by scientists at Guangxi University in collaboration with the Guangxi Buffalo Research Institute and international partners, performed an integrated transcriptomic analysis on milk-derived RNA from Murrah buffaloes at mid-lactation. The animals were divided into two clearly separated groups based on lactation performance. High-yield buffaloes produced an average of roughly 3,804 kilograms of milk per lactation, while low-yield animals averaged about 1,928 kilograms, a difference of nearly a factor of two. By sequencing the complete RNA content of milk from both groups, the researchers set out to identify the transcriptional features that accompany, and possibly underpin, this dramatic divergence in productivity.

The choice of milk as the sampling medium is one of the study’s most practical innovations. Mammary epithelial cells and other milk-associated cells are shed continuously into milk, meaning a simple sample of the fluid carries a molecular record of the secretory activity inside the udder. Unlike a mammary gland biopsy, milk collection is non-invasive, repeatable, and stress-free for the animal, which matters both for animal welfare and for the feasibility of scaling such measurements to breeding programs. The study’s authors note that no animals were euthanised, anaesthetised, or subjected to invasive procedures, and that the work was approved by the Animal Experiment Ethics Review Committee of Guangxi University.

Sequencing the milk RNA revealed a remarkably complex transcriptomic landscape. In total, the team identified 17,735 messenger RNAs, the protein-coding workhorses of the cell, alongside 2,641 long non-coding RNAs, 6,668 circular RNAs, and 578 microRNAs. This catalogue of four distinct RNA classes is what makes the study integrated rather than a conventional single-layer gene expression analysis. Long non-coding RNAs and circular RNAs can act as molecular sponges that bind microRNAs, thereby modulating how much protein gets produced from messenger RNA transcripts, a regulatory architecture known as the competing endogenous RNA, or ceRNA, network. Capturing all of these layers simultaneously from the same samples allowed the researchers to begin mapping how they interact.

The comparative analysis between high-yield and low-yield animals produced a coherent biological story. Buffaloes with higher milk output showed coordinated upregulation of genes involved in oxidative phosphorylation, the mitochondrial process that converts nutrients into cellular energy in the form of ATP. Genes tied to glucose metabolism, cellular metabolism in general, and protein synthesis were also elevated in the high-yield group. Taken together, these patterns point to an enhanced bioenergetic and biosynthetic state in the milk-associated cells of productive animals. In essence, the secretory cells of high-yielding buffaloes appear to run a hotter metabolic engine, burning more glucose through their mitochondria and building more protein, which is exactly what one would expect from tissue tasked with secreting roughly twice the volume of milk.

Equally telling was what was turned down. Transcripts associated with signal transduction, transporter activity, and apoptosis-related processes showed reduced expression in the high-yield group. Lower expression of apoptosis-related genes suggests that the milk cells of productive animals may resist programmed cell death, potentially sustaining secretory activity for longer. Reduced signalling and transporter transcripts hint at a shift in cellular priorities: rather than communicating with their environment or shuttling molecules across membranes, the cells of high-yield animals appear to channel resources into synthesis and secretion. The researchers are careful to frame these as associations rather than proven causes, describing the work as an exploratory framework, but the consistency of the metabolic signature across functional categories lends it weight.

By integrating the messenger RNA data with the non-coding RNA layers, the team constructed putative ceRNA networks that connect microRNAs, their messenger RNA targets, and the long non-coding and circular RNAs that sequester them. Within these networks, two candidate genes emerged as potentially associated with milk yield-related cellular pathways: SNX25, which encodes sorting nexin 25, a protein involved in intracellular trafficking, and DHPR, which encodes dihydropyridine reductase, an enzyme in the metabolic recycling pathway. These genes now stand as priority targets for future functional validation, the experimental step needed to determine whether they merely correlate with high yield or actively contribute to it.

The implications for buffalo breeding are potentially significant. Traditional genetic selection for milk yield in buffalo has been slowed by long generation intervals, limited pedigree and genomic records in many producing countries, and the sheer difficulty of measuring lactation traits early in an animal’s life. A transcriptomic marker panel derived from milk, measurable without harming the animal and potentially early in a lactation, could complement genomic selection by providing a functional readout of the actual secretory machinery. If the metabolic and ceRNA signatures identified here hold up in larger cohorts and across breeds, they could help identify high-potential animals sooner and guide the development of genetic improvement strategies tailored to buffalo biology rather than borrowed wholesale from dairy cattle.

The study also contributes a substantial public resource to the emerging field of livestock transcriptomics. Cataloguing nearly 18,000 messenger RNAs and thousands of non-coding transcripts from buffalo milk provides a reference for other researchers studying mammary biology, lactation physiology, and even milk quality traits, since many of the same metabolic pathways that drive yield also influence fat and protein composition. The raw sequencing data have been deposited in the Sequence Read Archive, allowing independent verification and reuse. The work was supported by the Guangxi Key Research and Development Program and the National Natural Science Foundation of China, reflecting the strategic importance of buffalo dairy production in southern China and across South Asia.

As with any exploratory transcriptomic study, caveats remain. The findings describe molecular features associated with divergent yield phenotypes at a single lactation stage in one breed, and the ceRNA networks are computational predictions that require experimental confirmation in mammary cells. The authors themselves position the work as a framework and a source of candidate targets rather than a definitive mechanism. Even so, the study marks a step toward a future in which a routine milk sample, already collected on every dairy farm every day, doubles as a molecular diagnostic window into the animal’s productive capacity. For an industry built on the biology of secretion, reading the transcriptome in the milk itself may prove to be one of the most elegant shortcuts breeding science has found in decades.

Subject of Research: Transcriptomic analysis of milk-derived RNA linked to divergent milk yield in buffalo

Article Title: Integrated transcriptomic analysis of milk-derived RNA associated with divergent milk yield phenotypes in buffalo

Article References: Integrated transcriptomic analysis of milk-derived RNA associated with divergent milk yield phenotypes in buffalo. (n.d.). https://doi.org/10.1186/s12864-026-13363-w

Image Credits: AI Generated

DOI: 10.1186/s12864-026-13363-w

Keywords: buffalo, milk yield, transcriptomics, RNA sequencing, non-coding RNAs, ceRNA network, oxidative phosphorylation, lactation, molecular breeding, gene expression, Murrah buffalo, BMC Genomics

Circular RNAs are an unusual class of molecules. Unlike ordinary linear messenger RNAs, which are transcribed, translated, and degraded in a straightforward sequence, circRNAs are formed when the splicing machinery of the cell joins the downstream end of an RNA molecule back to its upstream end, a process known as back-splicing. The result is a covalently closed loop with no free ends, which makes the molecule remarkably resistant to degradation by the exonucleases that normally chew up RNA. First noticed decades ago as rare splicing accidents, circRNAs are now recognized as abundant, conserved, and often functional components of the transcriptomes of animals and plants alike, with roles that include sponging microRNAs, modulating transcription, and in some cases even serving as templates for translation.

In the new work, Behzad Hajieghrari of Jahrom University and Mousa Torabi Giglou of the University of Mohaghegh Ardabili in Iran systematically reanalyzed strand-specific RNA sequencing data from rice exposed to simultaneous heat and drought stress. Because circRNAs lack the poly-A tails and defined ends of linear transcripts, detecting them requires specialized computational approaches. The researchers subjected the sequencing reads to rigorous quality control, mapped them to the rice genome, and then ran two independent circRNA prediction algorithms, CIRI2 and CIRCexplorer2, in parallel. Only candidates supported by both methods were retained, a dual-algorithm strategy designed to filter out false positives arising from repetitive sequence or misaligned reads. The screen yielded 208 high-confidence circRNAs distributed across all twelve rice chromosomes.

The comparative profiling revealed a striking pattern of stress-dependent circularization. Eighty-three circRNAs were detected exclusively in unstressed control samples, fifty-one appeared only in stressed samples, and seventy-four were shared between the two conditions. In other words, the circular transcriptome is not a static backdrop; it is remodeled when the plant senses combined stress. Junction-read analysis, which counts the sequencing reads that span the diagnostic back-splice junction, exposed a spectrum of circularization strength, from highly abundant circRNAs whose junction reads dominate their genomic loci to low-confidence candidates whose signals are partially masked by the background of linear transcripts from the same genes.

Genomic annotation showed that most of the identified circRNAs originated from exonic regions, with a substantial contribution from intergenic regions, and displayed a pronounced bias toward the negative DNA strand. Several host genes produced multiple distinct circRNA isoforms through alternative back-splicing, meaning that a single gene can generate a small family of circular molecules with potentially different regulatory partners. This isoform diversity adds a layer of complexity to the rice transcriptome that linear RNA analysis alone cannot capture, and it hints that alternative circularization may itself be a regulated process that the plant tunes under stress.

To infer what these circRNAs might be doing, the team performed functional enrichment analysis on their host genes. The results pointed to involvement in protein folding, nutrient reservoir activity, RNA degradation, and branched-chain amino acid catabolism. Each of these categories makes biological sense in the context of combined heat and drought. Protein folding machinery, including heat shock proteins, is central to surviving thermal damage; nutrient reservoir proteins reflect the metabolic reallocation that stress demands; RNA degradation pathways govern how quickly stress transcripts turn over; and amino acid catabolism connects to nitrogen mobilization and osmotic adjustment. The enrichment pattern suggests that circRNAs are not random byproducts but are embedded in the metabolic and proteostatic circuits that determine whether a rice plant tolerates or succumbs to compound stress.

Differential expression analysis between control and stressed libraries identified seven circRNAs specifically induced under combined heat and drought conditions. These stress-responsive candidates represent the most direct leads for future experimental work, since their induction implies that the plant actively upregulates their production as part of its adaptive program. Whether they act by sequestering microRNAs, interacting with RNA-binding proteins, or influencing the splicing of their own host genes remains to be determined, but their stress-specific behavior marks them as priority targets for functional validation.

The most intriguing part of the study concerns the predicted regulatory network built around these molecules. Drawing on the competing endogenous RNA, or ceRNA, hypothesis, the researchers predicted which microRNAs could bind each circRNA and which messenger RNAs those microRNAs could in turn regulate. In the ceRNA framework, a circRNA with binding sites for a particular microRNA can act as a molecular sponge, soaking up that microRNA and thereby relieving repression of the microRNA’s genuine mRNA targets. Network topology analysis of the resulting three-layer circRNA-microRNA-mRNA circuit pinpointed several microRNAs, including osa-miR414, osa-miR1439, and osa-miR2919, as candidate topological hubs, meaning they occupy central positions with many connections and could exert outsized influence over the network’s behavior. The predicted targets of these hub microRNAs encode stress-responsive transcription factors and signaling proteins, suggesting a plausible route by which circRNA abundance changes could ripple outward to reshape the expression of entire stress-response gene programs.

The authors are careful to frame these network findings as predictive. The ceRNA relationships were inferred computationally rather than demonstrated experimentally, and microRNA target prediction in plants, while reasonably accurate due to near-perfect base pairing requirements, still generates false positives. Nevertheless, the study delivers the first comprehensive map of circRNAs in rice under combined heat and drought stress, and it does so with a methodological transparency that should make follow-up work straightforward: the full lists of predicted circRNAs, their genomic coordinates, sequences, host gene annotations, differential expression results, and predicted interaction networks are all provided in supplementary data files. The researchers also note that the work received no external funding and was carried out with resources covered by the authors themselves.

The broader significance lies in what this means for crop improvement. Extreme weather events increasingly combine heat and water deficit during the rice growing season, and breeding for tolerance to each stress individually has not reliably produced varieties that withstand the combination. If circRNA-mediated regulation proves to be a genuine coordinating mechanism, it opens a new class of molecular markers and, eventually, engineering targets: circRNAs or the splicing elements that control their production could be tuned to bolster the plant’s post-transcriptional defenses. For now, the seven stress-induced circRNAs and the hub microRNAs osa-miR414, osa-miR1439, and osa-miR2919 constitute a concrete experimental agenda. Validating their interactions, confirming their sponging activity, and testing their effects on stress tolerance in living rice plants will determine whether these molecular rings are merely correlates of stress or true architects of the crop’s resilience.

Subject of Research: Circular RNA-mediated post-transcriptional regulation in rice under combined heat and drought stress

Article Title: Circular RNAs orchestrate integrated post-transcriptional responses to combined heat and drought stress in rice

Article References: Hajieghrari, B., & Giglou, M. T. (2026). Circular RNAs orchestrate integrated post-transcriptional responses to combined heat and drought stress in rice. Molecular Genetics and Genomics, 301(1), Article 183. https://doi.org/10.1007/s00438-026-02516-x

Image Credits: AI Generated

DOI: 10.1007/s00438-026-02516-x

Keywords: circular RNAs, rice, heat stress, drought stress, ceRNA network, microRNAs, back-splicing, RNA-seq, Oryza sativa, post-transcriptional regulation, plant stress responses, Molecular Genetics and Genomics