Methodological Progress

Regarding accuracy assessment in RNA structure prediction, we’ve introduced a hierarchical complexity level. Initially, prediction models must capture the Watson-Crick base pairs accurately, laying down the primary scaffolding for RNA secondary structure. Subsequent evaluation benchmarks consider the intricacies at helical junctions, such as 2-way, 3-way, and 4-way junctions. This demands algorithms to be adept not just at the secondary structural tier but also in ensuring global folding precision at the tertiary level. Preliminary results indicate that most models adeptly predict the overall topology and Watson-Crick base pairs, but challenges persist in intricate 3D junctions and multiple junction points. Notably, some prediction methods' precision requires refinement, especially when handling RNAs with non waston crick base pairs and long-range interactions.