BACKGROUND
Accurate RT-qPCR normalization depends on choosing reference genes whose expression is stable across the tissues, ages, and conditions being studied. In the yak, the stomach is especially important because efficient nutritional assimilation and energy metabolism help the animal survive on the Qinghai-Tibetan Plateau. The yak stomach is composed of four compartments—rumen, reticulum, omasum, and abomasum—and its function changes across development from birth to adulthood. Despite the importance of the yak stomach for digestion and nutrient absorption, no prior study had established suitable reference genes for RT-qPCR normalization across the full growth cycle. To address this gap, the authors used transcriptome sequencing data and published literature to identify and validate candidate reference genes for use in yak stomach tissues over time.
METHODS
This was an experimental animal and molecular biology study using 15 Maiwa yaks (7 males and 8 females) from the same herd. Animals were sampled at 5 growth stages: 0 days, 20 days, 60 days, 15 months, and 3 years old (adult). For each age, three yaks were slaughtered, and tissues from all four stomach compartments were collected. Total RNA was extracted from rumen, reticulum, omasum, and abomasum samples, assessed by NanoDrop2000, and checked by 1% agarose gel electrophoresis before cDNA synthesis. Fifteen candidate reference genes (CRGs) were selected: 7 from prior RNA-seq data and 8 from previous literature. The 7 transcriptome-derived genes were RPS15, RPS23, YWHAZ, RPL13A, ACTB, RPS9, and GAPDH, chosen because they had FPKM > 100 and CV < 20%. The 8 literature-based genes were UXT, DBNDD2, DDX54, HMBS, PPP1R11, MRPS15, MRPL39, and TBP. RT-qPCR primers were designed with lengths of 20 ± 3 bases and amplicons of 100 to 150 bp, and specificity was verified by 2% agarose gel electrophoresis, melting curve analysis, PCR product sequencing, and BLAST comparison. RT-qPCR was run in triplicate using the LightCycler 96 System, and a five-point standard curve with five-fold cDNA dilutions was used to calculate amplification efficiency and R^2. Expression stability was assessed using geNorm, NormFinder, BestKeeper, and the comparative Cq method, with RefFinder used to integrate rankings. The authors then validated the best reference-gene combinations by measuring HMGCS2, a key ketogenic enzyme, and comparing normalized RT-qPCR results with RNA-seq expression patterns.
KEY RESULTS
RNA quality was acceptable for all samples, with 260/280 ratios ranging from 1.8 to 2.2. From the transcriptome screen, 80 CRGs were initially identified, and 15 were carried forward for validation. Primer performance was strong overall, with amplification efficiencies ranging from 91 to 109%, amplicon sizes from 100 to 286 bp, and R^2 values of not less than 0.99. Across all assays, Cq values ranged from 18.49 to 32.67. RPS23 had the highest expression level among the CRGs, with Cq = 20.18 ± 0.76, while PPP1R11 had the lowest expression level, with Cq = 30.59 ± 0.89.
Stability ranking differed somewhat by algorithm, but the overall pattern was consistent. In geNorm, RPS15 and DBNDD2 were the most stable genes, with the lowest M-value of 0.48, and RPL13A was the least stable, with an M-value of 0.74. In NormFinder, RPS15 again ranked highest, with the lowest stability value of 0.35, while RPL13A was the least stable with a value of 0.62. BestKeeper identified RPS15 as the most stable and YWHAZ as the least stable. The comparative Cq method found MRPL39 to be the most stable and YWHAZ the most unstable. RefFinder integrated these results and ranked the genes from most to least stable as follows: RPS15 > MRPL39 > RPS23 > DDX54 > DBNDD2 > GAPDH > TBP > RPL13A > MRPS15 > PPP1R11 > ACTB > HMBS > RPS9 > UXT > YWHAZ.
GeNorm pairwise variation analysis suggested that three reference genes were sufficient for normalization. Using the accepted cut-off of V = 0.15, the authors concluded that adding a fourth gene was unnecessary because the normalization factor did not meaningfully improve beyond the three-gene combination. Accordingly, the recommended reference-gene set was RPS15, MRPL39, and RPS23. Validation with HMGCS2 showed that normalization using these three genes produced expression patterns that corresponded well with RNA-seq across the rumen, reticulum, omasum, and abomasum at 0 d, 20 d, 60 d, 15 m, and adult. In contrast, normalization using the three least stable genes, RPS9, UXT, and YWHAZ, produced discrepant results and, in the rumen, created significant differences where RNA-seq and the stable-gene normalization did not show differences; specifically, HMGCS2 differences between 60 days and 15 months became significant with the unstable set at p < 0.05.
CLINICAL IMPLICATIONS
Although this is not a human clinical study, it has important practical implications for veterinary science, animal physiology, and ruminant nutrition research. By establishing RPS15, MRPL39, and RPS23 as validated reference genes for yak stomach tissues across development, the study provides a robust normalization strategy for future RT-qPCR work on digestion, nutrient metabolism, and stomach development. The findings also reinforce that reference-gene stability is context-specific and that commonly used genes such as ACTB, GAPDH, or YWHAZ should not be assumed to be reliable without validation. For researchers studying yak adaptation, growth, or feeding strategies, using these three genes should improve the accuracy and reproducibility of transcript-level results.