eSkip-Finder

eSkip-Finder predicts optimal antisense oligonucleotide (ASO) sequences for exon skipping to modulate pre-mRNA splicing and support therapeutic ASO design.

Key Features:

• Machine Learning-Driven Prediction: eSkip-Finder employs machine learning models to predict the efficacy of novel ASO sequences in inducing exon skipping.
• Training on Experimental Datasets: The model is trained on extensive experimental ASO exon-skipping datasets to account for parameters influencing exon skipping outcomes.
• Sequence Input and Design Parameters: The predictor accepts exon and intron sequences and ASO length as inputs to generate candidate ASO sequences.
• Validation Against In Vitro Data: Predictive accuracy has been validated through correlation with in vitro experimental results for sequences excluded from the training dataset.
• Searchable ASO Database: The resource includes a database of exon-skipping ASOs searchable by gene name, species, and exon number.
• Consideration of Sequence and Structural Features: The algorithm considers sequence-specific interactions and structural properties of ASOs when evaluating efficacy.

Scientific Applications:

• Therapeutic Development: Optimizing ASO sequences to enable exon skipping for restoration of functional protein production in genetic disease research.
• Experimental Design and Prioritization: Designing experiments and prioritizing ASO candidates to increase predictive success and reduce empirical trial-and-error.

Methodology:

eSkip-Finder applies machine learning techniques trained on experimental ASO exon-skipping datasets, uses sequence-specific interaction and ASO structural features as model inputs, and assesses predictions by correlation with in vitro experimental results.