BaRDIC: robust peak calling for RNA-DNA interaction data

Area of Science:

• Genomics
• Molecular Biology
• Bioinformatics

Background:

• Non-coding RNAs (ncRNAs) are crucial regulators of cellular processes, interacting with specific genomic loci.
• Genome-wide sequencing (NGS) methods like 'one-to-all' and 'all-to-all' detect these RNA-DNA interactions but are susceptible to biases and noise.
• Accurate identification of specific RNA-DNA interaction 'peaks' is essential for understanding ncRNA function.

Purpose of the Study:

• To develop a novel computational tool, BaRDIC (Binomial RNA-DNA Interaction Caller), for robust peak detection in RNA-DNA interaction data.
• To address inherent biases in NGS data, including chromatin heterogeneity and distance-dependent decay of interaction frequency.
• To improve the accuracy and reliability of identifying genomic targets of non-coding RNAs.

Main Methods:

• BaRDIC employs a binomial model tailored for RNA-DNA interaction data.
• It uniquely accounts for chromatin heterogeneity and distance-dependent decay of interaction frequency.
• The method adaptively determines peak sizes based on individual RNA abundances and contact patterns.

Main Results:

• BaRDIC demonstrates superior performance in peak calling compared to existing algorithms.
• It effectively handles the sparsity characteristic of 'all-to-all' RNA-DNA interaction datasets.
• The tool provides more robust predictions of RNA-genomic target interactions.

Conclusions:

• BaRDIC represents a significant advancement in analyzing RNA-DNA interaction data.
• It offers a more accurate and reliable method for identifying ncRNA genomic targets.
• The freely available BaRDIC package facilitates further research in ncRNA function and regulation.