Table 1. Different MR strategies for nucleic acids of different lengths.

	Short	Medium	Long
Size (nt)	<30–40	40–200	>200
Typical secondary/tertiary structure	Hairpins	Combinations of hairpins	Complex tertiary structures
Availability of experimental models† (%)	83.5	13.1	3.4
Identification of structural homologues	Based on size, independently of sequence	By structure-similarity algorithm (sequence covariation)	By structure-similarity algorithm (sequence covariation)
Strategies to improve MR success	Generally unnecessary	Pruning bases/bases and sugars	Pruning bases/bases and sugars
		Deletion of loops and junctions	Using only selected domains
			Supporting the MR search using preliminary experimental phases
Limitations of MR using experimental models	Internal helical symmetry	R.m.s.d.‡ up to 2–3 Å	R.m.s.d.‡ up to 2–3 Å
Reference models in the absence of experimental data	Ideal helices modeled manually	Combinations of ideal helices modeled manually or	Combinations of ideal helices modeled manually (rare) or
		Three-dimensional motifs modeled de novo or	Three-dimensional motifs modeled de novo or
		Homology models	Homology models
Limitations of MR using in silico designed models	Internal helical symmetry	Difficulty in assigning small helical domains	Difficulty in assigning small helical domains
		R.m.s.d.‡ up to 1–1.5 Å	R.m.s.d.‡ up to 1–1.5 Å
References	Baikalov & Dickerson (1998 ▶)	Scott (2012 ▶)	Humphris-Narayanan & Pyle (2012 ▶), Marcia & Pyle (2012 ▶) and this work

^†Indicates the percentage of X-ray structures of nucleic acids of the corresponding size (statistics drawn from the PDB on 19 February 2013).

^‡Root-mean-square deviation between the MR search model and the target structure.