PPred-PCKSM: A multi-layer predictor for identifying promoter and its variants using position based features.

Area of Science:

• Genomics and Molecular Biology
• Bioinformatics
• Computational Biology

Background:

• Promoters are crucial DNA regions initiating gene transcription via RNA polymerase binding.
• Sigma factors in Escherichia coli (E.coli) direct RNA polymerase to specific promoter sequences.
• Existing promoter prediction methods require enhancement for improved accuracy and classification.

Purpose of the Study:

• To introduce PPred-PCKSM, a new multi-layer predictor for identifying E.coli promoters.
• To accurately classify six distinct types of E.coli sigma factors (σ70, σ24, σ28, σ32, σ38, σ54).
• To improve upon current state-of-the-art promoter prediction techniques.

Main Methods:

• Utilized a position-correlation based k-mer scoring matrix (PCKSM) for feature extraction.
• Employed concatenated feature sets from trimers and tetramers.
• Applied an artificial neural network (ANN) for the final prediction and classification task.
• Validated the model using 5-fold cross-validation on a benchmark dataset.

Main Results:

• Achieved a promoter prediction accuracy of 98.02%.
• Obtained a Matthews correlation coefficient (MCC) of 96.04% for promoter prediction.
• Demonstrated superior performance compared to existing state-of-the-art methods for E.coli promoter identification.

Conclusions:

• PPred-PCKSM effectively predicts promoters and their specific types in E.coli.
• The PCKSM feature extraction strategy combined with ANN significantly enhances prediction accuracy.
• This study offers a robust tool for advancing research in bacterial gene regulation.