 RNA base-pair geometry

RNA base-pair geometry specifies how two nucleobases contact and orient relative to one another, beyond their letter identities. It distinguishes different base pairing arrangements, including noncanonical base pairs, that can organize RNA secondary structure and three-dimensional folding.

A primary classification uses the contacting RNA base-pair edge of each partner: Watson–Crick, Hoogsteen or sugar edge. It also distinguishes cis and trans orientations of the two glycosidic bonds relative to the interbase hydrogen bonds. These are pairwise geometric descriptors; they are not simply labels for parallel versus antiparallel nucleic acid strands.^[ref]

An original counting explanation clarifies the twelve basic families. Ignoring exchange of the two partner labels, three edge types allow three same-edge combinations and three mixed-edge combinations: six possibilities. Combining each with cis or trans orientation gives $6\times 2=12$. This counts contact-geometry families, not twelve possible nucleotide sequences or twelve equally common structures.

For example, knowing only that a contact is guanine–adenine does not specify which RNA base-pair edges participate. Replacing that pair by another two-base combination also does not guarantee the new contact fits the same surrounding backbone. Geometry can therefore explain why one sequence substitution preserves a structural motif while another disrupts it. The classification was designed to support structural comparison and recognition of geometrically similar pairs.^[ref]

Experimental coordinates or solution restraints are needed to assign an actual contact. A proposed hydrogen bond requires a compatible arrangement of donor and acceptor groups, not merely a favorable pair of letters. A secondary-structure arc records a relationship between residues; a full RNA base-pair geometry adds spatial information that the arc alone cannot encode.