Lecture 4
Comparative analysis of RNA tertiary structure
	Principle
	RNA folding - 2¡ first, then 3¡
		comparison of helical stacks with protein subdomains (not according to Moore)
	Data requirements
		# of seqs, range and sampling of variation
		usually from database mining or natural pops
	Structural interpretation
		the interaction not always easily interpretable
		basis for local area 3D models
		ID interesting interaction for physical study
	Statistical analysis of tertiary contacts
		in principle no different than for secondary structure
		lone basepairs - e.g. rRNA lone pair, M.fe P lone pair
		pseudoknots
			are pseudoknots really tertiary? which helix? show P figure
		non-W/C basepairs
			not just GU or GA:GA's that covary with WC bps
			Archaeal P P17
			P P11
		tetraloops - rGNRAr/yUNCGr/rCUUGy
			sheared pair
			if conserved, look for docking site:
			GNRA tetraloop-receptor docking - specify covarying nts
			GAAA tetraloop-receptor docking - covariation of their presence (seqs invariant)
		other motifs
	Insertions/deletions & RNA substructures
		P P19
			discrete structural element
			show that the subset that have the insertion have the covariation that shows bps
		P P16/17
			discrete structural element
			those than have them also have P6 - others w/ P 5.1, P15.1
		3D model
	Helical stacks
		P P13/14
			secondart structures
			3D model
		P P5/15
	Secondary structure motif substitutions
		interpretation of correlations - isopair
		P P17/6 vs P5.1
		P P10.1 vs P13/14
		3D model
		evolutionary series of type A->B
	Strengths
		same as for secondary 
		often can ID interactions that are genetically or biochemically transparent
		can ID interesting motifs for physical analysis
	Weaknesses
		structural interpretation not always easy unless it's a known motif
			example - P12:P13 in P
		not as well correlated as secondary interactions - harder to ID, so more seqs needed
		usually highly conserved seqs - therefore more seqs needed
		harder to confirm genetically or biochemically (but use X-linking)