--------   pknotsRG   -------
 
Folding canonical RNA secondary structures including pseudoknots
----------------------------------------------------------------
Robert Giegerich      ([email protected])
Jens Reeder	      ([email protected])	

Practical Computer Science
Faculty of Technology
University of Bielefeld
----------------------------------------------------------------

pknotsRG is a tool for thermodynamic folding of RNA secondary
structures, including the class of canonical simple recursive
pseudoknots. This class and the algorithms are described in detail
in:

"Design, Implementation and Evaluation of a Practical Pseudoknot 
Folding Algorithm based on Thermodynamics",
Jens Reeder and Robert Giegerich,
BMC Bioinformatics 2004.

----------------------------------------------------------------

Files:  Algebras.lhs
	Energy.lhs
	Foldingspace.lhs
	RNACombinators.lhs
	Intloop.lhs
	Intloop21.lhs
	Intloop22.lhs
	pknotsRG-mfe.lhs
	pknotsRG-enf.lhs
	pknotsRG-loc.lhs
	ADPfold.lhs 	(an RNAfold clone using ADP)
	
Building:
	We compile the files with ghc using make (see makefile for details). First try
	
		make
		
	If this doesn't work, you have to run ghc by yourself:		

		ghc --make -O2 pknotsRG-mfe.lhs
	
----------------------------------------------------------------

Currently there are three different version available:

	pknotsRG-mfe:  	computes the best structure (i.e. the
			structure with minimum free energy).
	pknotsRG-enf:  	computes the best structure that actually
			contains at least one pseudoknot.
	pknotsRG-loc:   computes the best "compact" pseudoknot,
			i.e. the structure with the lowest energy
			to length ratio.


Usage: 
		pknotsRG-mfe ucaaguauuccgaagcucaacgggaaaaugagcua
        or
		pknotsRG-mfe AUCUGUCAUCUAUUGCUAUCU

	For the other variants substitute mfe with enf or loc.
	Optional the size of the pseudoknot can be restricted.
	This accelerates the computation substantially. We used 
	this parameter to search for pseudoknots of maxsize 100
	within sequences of length 1000.
	
		pknotsRG-mfe sequence optional_maxsize


	Note: All non ACGTU characters are internally mapped to N
	      and are disallowed in any basepairings.



Output:
	The output consist of the input sequence, the secondary
	structure, ad the minimum free energy.
	Basepairs of the first pseudoknot helix are denoted by
	'{' and '}', the second helix by '[' and ']' . Basepairs
	not involved in a pseudoknot correspond to normal brackets,
	 '(' and ')'.

	Example:
	
		UCAAGUAUUCCGAAGCUCAACGGGAAAAUGAGCUA
		.......[[[[[.{{{{{{.]]]]]...}}}}}}. ( -14.9)
	
	Since pknotsRG-loc finds the best local pseudoknot, the
	output of this variant contains also the start and end position
	of the corresponding subsequence

		7........................34
		UUCCGAAGCUCAACGGGAAAAUGAGCU
		[[[[[.{{{{{{.]]]]]...}}}}}} (-14.51)


	
-----------------------------------------------------------------

The energy model we use for pseudoknots:

Destabilizing:
	creating a new pseudoknot :		9.0
	basepair inside pseudoknot:	        0.0
	not paired base in pk     :		0.3

Stabilizing:
	stacking of basepairs     :     	stack
	base dangling of a pk pair:     	dangle
	coaxial stacking          :     	stack

all values in (kcal/mol)


stack and dangle are the energies for nested structures like in
mfold-3.1. If any nested or unnested structures occur inside a
pseudoknot, their energy, of course, contributes to the overall
pseudoknot energy.