Lecture 2
Thermodynamic prediction and chemical/enzymatic probing of RNA structure
	Thermodynamic prediction
		MFOLD
			algorithm
				dynamic algorithm
				energy matrix vs traceback
				bp scoring via n.n. Turner rules
				other factors - loop size, tetraloops, etc
			parameters
				DNA vs RNA
				circular vs linear
				temperature
				max number of sequences to examine
				p - how different must they be?
				energy cut-off
				bp specifications/exclusions
			output files
				scoring matrix
				connect files
				graphics files
				colorization by frequency of occurance
		When this approach is useful
			small (<100) RNAs or discrete substructures
			as a source of reasonable structures for beginning CSA
			w/ probing where you only have a single sequence
			works best with thermophilic RNAs
			unalignable SELEX products
		examples
			D.radiodurans P
			D.radiodurans P P8
			E1 RNA
		weaknesses
			long sequences
			pseudoknots
			large loops, esp. internal loops
			details
			high A+U or G+C sequences
	Chemical & enzymatic probing
		principle
		common probes (TABLE)
			enzymes
				RNase U2 - unpaired G
				RNase T1 - unpaired A>G>>C>U
				S1 nuclease - unpaired N
				RNase V1 - stacked (paired?) N
			chemicals
				DMS (dimethylsulfate) - A(N1), C(N3), G(N7)
				CMCT - G(N1), U(N3)
				DEPC (diethylpyrocarbonate) - A(N7)
				kethoxal - G(N1), G(N2)
				ENU (ethylnitrosourea) - any accessible P
				Fe-EDTA (hydroxide radicals) - ribose
				Pb++ - ribose-P in metal-binding sites
		examples
			from paper
			DMS footprints from Connie Yarian
		weaknesses
			MUST use at <1 cleavage per molecule
			steric hinderance
			specificities not very defined
			interpretation difficult
		strengths
			physiological conditions
			best at identifying CHANGES in structure or footprints
			can be made quantitative, e.g. to derive binding constants
			can be used with homologous seqs from different organisms to ID important results
	Combination for prediction of structure from a single or a few sequences
		example - E1 RNA, pVX virus 5' end