Long noncoding RNAs (lncRNAs) play central roles in coordinating gene expression, yet their structure-function relationships are poorly understood. Most lncRNAs control spatiotemporal patterns of development, direct epigenetic changes, and scaffold phase-separated domains that act as hubs for metabolic and genetic processes. Although some lncRNA functions are mediated through sequence-specific contacts with DNA, they also fold to form complex three-dimensional structures, driving interactions with proteins. There is considerable evidence that lncRNAs recruit specific RNA binding proteins (RBPs) through repeated structural modules, tailoring their cellular effects by their selection of modules during splicing. This is supported by the finding that lncRNAs are highly alternatively spliced, have independently folding exons, frequently include transposable elements, and feature a high degree of structural conservation.
Here, we present our work to interrogate lncRNA structural modules, initially within chromosome 21. We use ECSFinder to identify evolutionarily conserved RNA secondary structures (ECSs) in mammalian genome alignments. This conservation demonstrates that lncRNA structure, rather than sequence, is likely functional. Next, we create a homology map of RNA structures similar to those identified ECSs across human chromosome 21, and analyse their locations. Following this, we plan to use eCLIP-seq data to identify putative protein partners of each module, by enrichment for the genomic positions of RNAs bound by specific RBPs in the vicinity of homologous RNA structures. This approach will provide an avenue to develop an expanded Rfam database of structure-function relationships in lncRNAs and their isoforms, and thus, help to illuminate their cellular roles.