PromoterPredict

PromoterPredict predicts σ70 promoter strength from DNA sequence by modeling the logarithm of promoter strength as a weighted sum of sequence profile scores at the –10 and –35 regions.

Key Features:

• Sequence-Based Modelling: Trains on a well-characterized set of promoters and models a significant linear relationship between the logarithm of promoter strength and a weighted sum of sequence profile scores at the –10 and –35 regions.
• Dynamic Learning Capability: Supports iterative refinement of model parameters using user-supplied data to improve prediction accuracy and confidence.
• Equal Contribution of Sequence Regions: Analysis indicates the –10 and –35 hexamer sequences contribute nearly equally to determining promoter strength.

Scientific Applications:

• Genetic engineering: Inform selection and modification of σ70 promoters for engineered gene expression constructs.
• Synthetic biology: Aid design and optimization of gene expression systems by predicting promoter strength from sequence.
• Promoter class generalization: Provide a framework that can be evaluated for applicability to promoter classes beyond σ70 across bacterial species.

Methodology:

A multivariate linear regression model is trained on a dataset of well-characterized σ70 promoters; the model predicts promoter strength via the logarithmic dependence on sequence profile scores from the –10 and –35 regions and can be retrained with user-supplied data to refine parameters.