Inferring mammalian tissue-specific regulatory conservation by predicting tissue-specific differences in open...

Area of Science:

• Genomics
• Computational Biology
• Evolutionary Biology

Background:

• Evolutionary conservation aids in identifying functionally significant genomic regions.
• Traditional nucleotide alignment methods struggle to capture conservation in enhancers, which rely on complex regulatory codes.
• Enhancer function is governed by specific combinations of regulatory elements, allowing for conservation despite high nucleotide turnover.

Purpose of the Study:

• To develop a novel machine learning approach for evaluating enhancer conservation based on regulatory sequence codes.
• To predict tissue-specific enhancer activity and identify conserved or divergent regulatory functions across species.
• To provide a framework for annotating regulatory function and studying enhancer evolution in a large-scale genomic context.

Main Methods:

• Trained a convolutional neural network to predict tissue-specific open chromatin (a proxy for enhancer activity) across mammals.
• Applied the model to distinguish conserved versus lost regulatory activity based on genomic sequence.
• Systematically evaluated model performance and compared it against nucleotide alignment-based approaches.

Main Results:

• The machine learning model accurately predicts tissue-specific conservation and divergence in open chromatin between primate and rodent species.
• The approach significantly outperforms leading nucleotide alignment-based methods in evaluating enhancer conservation.
• Brain enhancers associated with neuron activity show a greater tendency for predicted lineage-specific open chromatin compared to other enhancers.

Conclusions:

• The developed framework enables annotation of tissue-specific regulatory function across numerous genomes.
• It facilitates the study of enhancer evolution by analyzing predicted regulatory differences rather than solely relying on nucleotide-level conservation.
• This method offers a new perspective on understanding the evolutionary dynamics of gene regulation.