Computational prediction of human deep intronic variation.

Area of Science:

• Genomics
• Bioinformatics
• Computational Biology

Background:

• Whole-genome sequencing enables detection of genetic variations in deep intronic regions.
• Distinguishing functionally relevant variants from non-effect variants computationally is challenging due to limited benchmarks.

Purpose of the Study:

• To provide an overview of computational methods for analyzing deep intronic variation.
• To evaluate the performance of these tools across different intronic regions and molecular mechanisms.

Main Methods:

• Leveraged diverse datasets to evaluate tool performance.
• Compared SpliceAI with newer methods extending its implementation.
• Devised a quantitative assessment for tool interpretability.

Main Results:

• Tool performance varied significantly depending on the intronic region and variant mechanism.
• Cryptic splice site variants were better predicted than those affecting regulatory elements.
• Interpretable tools showed decreased predictive power compared to black-box methods.

Conclusions:

• Findings offer practical recommendations for computational tool usage in deep intronic regions.
• Provides a reference framework for informed decision-making in variant analysis.