The identification of cis-regulatory elements: A review from a machine learning perspective

Area of Science:

• Genomics
• Bioinformatics
• Computational Biology

Background:

• Non-coding DNA, once termed 'junk DNA', harbors critical cis-regulatory elements like promoters and enhancers.
• These elements control gene expression across various cell types, conditions, and developmental stages.
• Disruptions in cis-regulatory regions can lead to significant phenotype changes.

Purpose of the Study:

• To review machine learning (ML) approaches for predicting cis-regulatory elements, including transcription factor binding sites, enhancers, and promoters.
• To highlight the importance of ML in analyzing complex genomic data from next-generation sequencing (NGS).
• To encourage computational experts and data scientists to contribute to advancing this field.

Main Methods:

• Focus on machine learning techniques applied to next-generation sequencing data.
• Discuss methods for predicting transcription factor binding sites, enhancers, and promoters.
• Provide data sources to facilitate the testing and development of novel prediction methods.

Main Results:

• Machine learning offers accurate and efficient computational solutions for identifying cis-regulatory elements.
• NGS technologies provide in-depth genomic feature capture essential for these analyses.
• The review consolidates current ML strategies for regulatory element prediction.

Conclusions:

• Precise identification of regulatory elements is vital for understanding transcriptional regulation and its clinical implications.
• ML is indispensable for navigating the complexity of cis-regulatory events and large-scale sequencing data.
• Further advancements in ML are needed to fully decipher the functional landscape of the human genome.