Evaluating Methods for the Prediction of Cell Type-Specific Enhancers in the Mammalian Cortex

Area of Science:

• Neuroscience
• Genomics
• Computational Biology

Background:

• Identifying cell type-specific enhancers is crucial for developing genetic tools to study the mammalian brain.
• Computational enhancer prediction methods have been validated in fruit flies but not yet in mammalian brains.

Purpose of the Study:

• To assess machine learning and feature-based methods for predicting functional cell type-specific enhancers in the mouse cortex.
• To establish a benchmark for enhancer prioritization algorithms in mammalian brain research.

Main Methods:

• The Brain Initiative Cell Census Network (BICCN) Challenge organized the evaluation of computational methods.
• Methods were assessed using in vivo validation data from hundreds of validated cortical cell type-specific enhancers injected into mice.
• Machine learning and feature-based approaches were compared for their ability to nominate enhancer DNA sequences.

Main Results:

• Open chromatin regions were identified as a key predictor of functional enhancers.
• Sequence models improved the prediction of non-functional enhancers, allowing for deprioritization.
• Sequence models revealed cell type-specific transcription factor codes for in silico enhancer design.

Conclusions:

• This challenge established a benchmark for enhancer prioritization algorithms in mammalian cortical cell types.
• Computational approaches and molecular information are crucial for identifying functional enhancers.
• The findings advance the understanding of the mammalian cortex's gene regulatory landscape and the design of targeted genetic tools.