Follow us
P147 – Oscillating CpG Sites in Human Blood as Molecular Clocks for Estimating the Time-of-day of Blood Deposition
In forensic molecular biology, biological traces recovered from crime scenes are primarily used for donor identification. However, for evidence-based reconstruction of criminal events, contextual information on when a trace was deposited can be equally important. In particular, predicting the time-of-day of deposition (ToD) could provide valuable investigative information. Existing methods relying on diurnally rhythmic mRNA markers show potential for ToD prediction, but their vulnerable to degradation limit practical applicability. Here, we developed an epigenetic strategy for ToD estimation of blood using oscilating modified cytosines signatures. Building on evidence that a subset of cytosine methylation markers exhibits rhythmic variation across the daily cycle, we investigated whether such oscillatory epigenetic markers can be used for forensic blood deposition time estimation within one day. Blood samples collected at defined time points across the day were subjected to DNA methylation profiling to identify candidate prediction markers which were then selected and integrated into predictive models for ToD estimation. We performed epigenome-wide screening using the Illumina Infinium MethylationEPIC v2.0 BeadChip on peripheral whole blood from 10 healthy subjects sampled every 4 h over a complete 24 h cycle (70 samples total). We comprehensively identified 3,635 CpGs exhibiting significant diurnal rhythmic oscillations (JTK method, BH-Q < 0.05) in human peripheral blood. The dataset was split into training set and test set based on individuals, and 5-fold cross-validation was used for evaluating model performance. By applying Random Forest machine learning for feature selection, we constructed an initial model using TimeSignatR algorithm. Based on the top 50 important features, this model achieved a mean absolute error (MAE) of 1.85 h (R2 = 0.64) in test set. In conclusion, we established a DNA methylation-based approach for blood ToD prediction and provide evidence that circadian epigenetic information can assist forensic trace contextualization.
Authors
- Junyan Wang (Hebei Medical University, China)
- Tianyi Song (Hebei Medical University, China)
- Li Yang (Hebei Medical University, China)
- Shujin Li (Hebei Medical University, China)
On the same topic
