iDHU-Ensem

iDHU-Ensem predicts dihydrouridine (D) modification sites in transfer RNA (tRNA) sequences to enable computational identification of tRNA dihydrouridine modifications relevant to tRNA folding and disease studies.

Key Features:

• D site prediction: Predicts dihydrouridine (D) sites within transfer RNA (tRNA) sequences.
• Feature extraction: Implements novel feature extraction mechanisms that transform biological sequences into vector representations for computational analysis.
• Ensemble learning: Integrates multiple algorithms using ensemble learning models.
• Stacking ensemble performance: The stacking ensemble achieved accuracy 0.98, specificity 0.98, sensitivity 0.97, and Matthews Correlation Coefficient (MCC) 0.92.
• Evaluation: Performance was assessed using k-fold cross-validation and independent testing procedures.
• Benchmarking: Benchmarking on an independent test set showed superior accuracy compared to existing predictors.

Scientific Applications:

• Dihydrouridine site identification: Identification and mapping of dihydrouridine modifications in tRNA sequences.
• tRNA structural studies: Investigation of tRNA folding and conformational flexibility associated with dihydrouridine.
• Disease association research: Research into links between dihydrouridine modifications and diseases such as lung cancer.

Methodology:

Novel feature extraction converting sequences to vectors; ensemble learning models including a stacking ensemble; evaluation via k-fold cross-validation and independent testing.