Development of Mathematical Models Using circRNA Combinations (circTulp4, circSlc8a1, and circStrn3) in Mouse Brain Tissue for Postmortem Interval Estimation

The postmortem interval (PMI) is defined as the time interval between physiological death and the examination of the corpse, playing a critical role in forensic investigations. Traditional PMI estimation methods are often influenced by subjective and environmental factors. Circular RNAs (circRNAs), known for their stability, abundance, and conservation in brain tissue, show promise as biomarkers for PMI estimation. However, research on circRNAs in this context remains limited. This study aimed to develop PMI estimation models using circRNAs across multiple temperatures. By employing semi-quantitative reverse transcription-PCR, circTulp4, circSlc8a1, and circStrn3 were identified as reliable biomarkers for mouse brain tissue. Mathematical models were constructed using the reference genes 28S rRNA, mt-co1, and circCDR1as. At 4 °C, most equations had p-values below 0.05, with the equation using circSlc8a1 as a marker exhibiting the highest goodness of fit. Validation results indicated that the equation using circTulp4 as the reference gene had the highest accuracy. When applying the combined aforementioned three circRNAs, the equation using circCDR1as as the reference gene showed better accuracy. At 25 °C, all equations had R2 values greater than 0.86, but most cubic equations had p-values above 0.05. Validation results demonstrated that the circTulp4/mt-co1 equation had the highest accuracy. When applying combined circRNAs, the R2 values improved, and long-term PMI estimation was more accurate than short-term PMI estimation. At 35 °C, the linear equations had significantly poorer goodness of fit compared to nonlinear equations, and nonlinear equations exhibited better accuracy than linear equations. When applying the combined aforementioned three circRNAs, the accuracy of the three reference genes was similar, and the accuracy of long-term PMI estimation was consistently higher than that of short-term estimation. For the three-dimensional models, all R2 values exceeded 0.75 with p-values significantly below 0.0001. Validation results demonstrated higher accuracy at 25 °C and 35 °C, with superior performance for long-term PMI estimation. In summary, this study constructed PMI estimation models under multiple temperature conditions based on highly expressed circRNAs in mouse brain tissue, highlighting circTulp4, circSlc8a1, and circStrn3 as novel biomarkers. These findings offer a complementary tool for PMI estimation, particularly for long-term PMI estimation.