Energy Parameters

For secondary structure free energy evaluation we usually utilize the set of thermodynamic Nearest Neighbor energy parameters also used in other software, such as UNAfold and RNAstructure.

Specialized Modules:

Available Parameter Sets

While the RNAlib already contains a compiled-in set of the latest Turner 2004 Free Energy Parameters, we defined a file format that allows to change these parameters at runtime. The ViennaRNA Package already comes with a set of parameter files containing

Turner 1999 RNA parameters
Mathews 1999 DNA parameters
Andronescu 2007 RNA parameters
Mathews 2004 DNA parameters

Energy Parameter API

Defines

VRNA_GQUAD_MAX_STACK_SIZE: #include <ViennaRNA/params/basic.h>

VRNA_GQUAD_MIN_STACK_SIZE: #include <ViennaRNA/params/basic.h>

VRNA_GQUAD_MAX_LINKER_LENGTH: #include <ViennaRNA/params/basic.h>

VRNA_GQUAD_MIN_LINKER_LENGTH: #include <ViennaRNA/params/basic.h>

VRNA_GQUAD_MIN_BOX_SIZE: #include <ViennaRNA/params/basic.h>

VRNA_GQUAD_MAX_BOX_SIZE: #include <ViennaRNA/params/basic.h>

Typedefs

typedef struct vrna_param_s vrna_param_t: #include <ViennaRNA/params/basic.h>
Typename for the free energy parameter data structure vrna_params.

typedef struct vrna_exp_param_s vrna_exp_param_t: #include <ViennaRNA/params/basic.h>
Typename for the Boltzmann factor data structure vrna_exp_params.

typedef struct vrna_param_s paramT

#include <ViennaRNA/params/basic.h>

Old typename of vrna_param_s.

Deprecated:: Use vrna_param_t instead!

typedef struct vrna_exp_param_s pf_paramT

#include <ViennaRNA/params/basic.h>

Old typename of vrna_exp_param_s.

Deprecated:: Use vrna_exp_param_t instead!

Functions

vrna_param_t *vrna_params(vrna_md_t *md)

#include <ViennaRNA/params/basic.h>

Get a data structure containing prescaled free energy parameters.

If a NULL pointer is passed for the model details parameter, the default model parameters are stored within the requested vrna_param_t structure.

Parameters:

md – A pointer to the model details to store inside the structure (Maybe NULL)

Returns:

A pointer to the memory location where the requested parameters are stored

vrna_param_t *vrna_params_copy(vrna_param_t *par)

#include <ViennaRNA/params/basic.h>

Get a copy of the provided free energy parameters.

If NULL is passed as parameter, a default set of energy parameters is created and returned.

See also

vrna_params(), vrna_param_t

Parameters:

par – The free energy parameters that are to be copied (Maybe NULL)

Returns:

A copy or a default set of the (provided) parameters

vrna_exp_param_t *vrna_exp_params(vrna_md_t *md)

#include <ViennaRNA/params/basic.h>

Get a data structure containing prescaled free energy parameters already transformed to Boltzmann factors.

This function returns a data structure that contains all necessary precomputed energy contributions for each type of loop.

In contrast to vrna_params(), the free energies within this data structure are stored as their Boltzmann factors, i.e.

\( exp(-E / kT) \)

where \( E \) is the free energy.

If a NULL pointer is passed for the model details parameter, the default model parameters are stored within the requested vrna_exp_param_t structure.

See also

vrna_md_t, vrna_md_set_default(), vrna_params(), vrna_rescale_pf_params()

Parameters:

md – A pointer to the model details to store inside the structure (Maybe NULL)

Returns:

A pointer to the memory location where the requested parameters are stored

vrna_exp_param_t *vrna_exp_params_comparative(unsigned int n_seq, vrna_md_t *md)

#include <ViennaRNA/params/basic.h>

Get a data structure containing prescaled free energy parameters already transformed to Boltzmann factors (alifold version)

If a NULL pointer is passed for the model details parameter, the default model parameters are stored within the requested vrna_exp_param_t structure.

Parameters:

n_seq – The number of sequences in the alignment
md – A pointer to the model details to store inside the structure (Maybe NULL)

Returns:

A pointer to the memory location where the requested parameters are stored

vrna_exp_param_t *vrna_exp_params_copy(vrna_exp_param_t *par)

#include <ViennaRNA/params/basic.h>

Get a copy of the provided free energy parameters (provided as Boltzmann factors)

If NULL is passed as parameter, a default set of energy parameters is created and returned.

See also

vrna_exp_params(), vrna_exp_param_t

Parameters:

par – The free energy parameters that are to be copied (Maybe NULL)

Returns:

A copy or a default set of the (provided) parameters

void vrna_params_subst(vrna_fold_compound_t *fc, vrna_param_t *par)

#include <ViennaRNA/params/basic.h>

Update/Reset energy parameters data structure within a vrna_fold_compound_t.

Passing NULL as second argument leads to a reset of the energy parameters within fc to their default values. Otherwise, the energy parameters provided will be copied over into fc.

SWIG Wrapper Notes:

This function is attached to vrna_fc_s objects as overloaded params_subst() method.

When no parameter is passed, the resulting action is the same as passing NULL as second parameter to vrna_params_subst(), i.e. resetting the parameters to the global defaults. See, e.g. RNA.fold_compound.params_subst() in the Python API.

Parameters:

fc – The vrna_fold_compound_t that is about to receive updated energy parameters
par – The energy parameters used to substitute those within fc (Maybe NULL)

void vrna_exp_params_subst(vrna_fold_compound_t *fc, vrna_exp_param_t *params)

#include <ViennaRNA/params/basic.h>

Update the energy parameters for subsequent partition function computations.

This function can be used to properly assign new energy parameters for partition function computations to a vrna_fold_compound_t. For this purpose, the data of the provided pointer params will be copied into fc and a recomputation of the partition function scaling factor is issued, if the pf_scale attribute of params is less than 1.0.

Passing NULL as second argument leads to a reset of the energy parameters within fc to their default values

SWIG Wrapper Notes:

This function is attached to vrna_fc_s objects as overloaded exp_params_subst() method.

When no parameter is passed, the resulting action is the same as passing NULL as second parameter to vrna_exp_params_subst(), i.e. resetting the parameters to the global defaults. See, e.g. RNA.fold_compound.exp_params_subst() in the Python API.

Parameters:

fc – The fold compound data structure
params – A pointer to the new energy parameters

void vrna_exp_params_rescale(vrna_fold_compound_t *fc, double *mfe)

#include <ViennaRNA/params/basic.h>

Rescale Boltzmann factors for partition function computations.

This function may be used to (automatically) rescale the Boltzmann factors used in partition function computations. Since partition functions over subsequences can easily become extremely large, the RNAlib internally rescales them to avoid numerical over- and/or underflow. Therefore, a proper scaling factor \(s\) needs to be chosen that in turn is then used to normalize the corresponding partition functions \(\hat{q}[i,j] = q[i,j] / s^{(j-i+1)}\).

This function provides two ways to automatically adjust the scaling factor.

Automatic guess
Automatic adjustment according to MFE

Passing NULL as second parameter activates the automatic guess mode. Here, the scaling factor is recomputed according to a mean free energy of 184.3*length

cal for random sequences.

On the other hand, if the MFE for a sequence is known, it can be used to recompute a more robust scaling factor, since it represents the lowest free energy of the entire ensemble of structures, i.e. the highest Boltzmann factor. To activate this second mode of

automatic adjustment according to MFE, a pointer to the MFE value needs to be passed as second argument. This value is then taken to compute the scaling factor as \( s = exp((sfact * MFE) / kT / length )\), where sfact is an additional scaling weight located in the vrna_md_t data structure of exp_params in fc.

The computed scaling factor \(s\) will be stored as pf_scale attribute of the exp_params data structure in fc.

SWIG Wrapper Notes:

This function is attached to vrna_fc_s objects as overloaded exp_params_rescale() method.

When no parameter is passed to this method, the resulting action is the same as passing NULL as second parameter to vrna_exp_params_rescale(), i.e. default scaling of the partition function. Passing an energy in kcal/mol, e.g. as retrieved by a previous call to the mfe() method, instructs all subsequent calls to scale the partition function accordingly. See, e.g. RNA.fold_compound.exp_params_rescale() in the Python API.

Note

This recomputation only takes place if the pf_scale attribute of the exp_params data structure contained in fc has a value below 1.0.

Parameters:

fc – The fold compound data structure
mfe – A pointer to the MFE (in kcal/mol) or NULL

void vrna_params_reset(vrna_fold_compound_t *fc, vrna_md_t *md)

#include <ViennaRNA/params/basic.h>

Reset free energy parameters within a vrna_fold_compound_t according to provided, or default model details.

This function allows one to rescale free energy parameters for subsequent structure prediction or evaluation according to a set of model details, e.g. temperature values. To do so, the caller provides either a pointer to a set of model details to be used for rescaling, or NULL if global default setting should be used.

SWIG Wrapper Notes:

This function is attached to vrna_fc_s objects as overloaded params_reset() method.

When no parameter is passed to this method, the resulting action is the same as passing NULL as second parameter to vrna_params_reset(), i.e. global default model settings are used. Passing an object of type vrna_md_s resets the fold compound according to the specifications stored within the vrna_md_s object. See, e.g. RNA.fold_compound.params_reset() in the Python API.

See also

vrna_exp_params_reset(), vrna_params_subs()

Parameters:

fc – The fold compound data structure
md – A pointer to the new model details (or NULL for reset to defaults)

void vrna_exp_params_reset(vrna_fold_compound_t *fc, vrna_md_t *md)

#include <ViennaRNA/params/basic.h>

Reset Boltzmann factors for partition function computations within a vrna_fold_compound_t according to provided, or default model details.

This function allows one to rescale Boltzmann factors for subsequent partition function computations according to a set of model details, e.g. temperature values. To do so, the caller provides either a pointer to a set of model details to be used for rescaling, or NULL if global default setting should be used.

SWIG Wrapper Notes:

This function is attached to vrna_fc_s objects as overloaded exp_params_reset() method.

When no parameter is passed to this method, the resulting action is the same as passing NULL as second parameter to vrna_exp_params_reset(), i.e. global default model settings are used. Passing an object of type vrna_md_s resets the fold compound according to the specifications stored within the vrna_md_s object. See, e.g. RNA.fold_compound.exp_params_reset() in the Python API.

Parameters:

fc – The fold compound data structure
md – A pointer to the new model details (or NULL for reset to defaults)

void vrna_params_prepare(vrna_fold_compound_t *fc, unsigned int options): #include <ViennaRNA/params/basic.h>

vrna_param_t *get_parameter_copy(vrna_param_t *par): #include <ViennaRNA/params/basic.h>

vrna_exp_param_t *get_scaled_pf_parameters(void)

#include <ViennaRNA/params/basic.h>

get a data structure of type vrna_exp_param_t which contains the Boltzmann weights of several energy parameters scaled according to the current temperature

Deprecated:: Use vrna_exp_params() instead!

Returns:: The data structure containing Boltzmann weights for use in partition function calculations

vrna_exp_param_t *get_boltzmann_factors(double temperature, double betaScale, vrna_md_t md, double pf_scale)

#include <ViennaRNA/params/basic.h>

Get precomputed Boltzmann factors of the loop type dependent energy contributions with independent thermodynamic temperature.

This function returns a data structure that contains all necessary precalculated Boltzmann factors for each loop type contribution.

In contrast to

get_scaled_pf_parameters(), this function enables setting of independent temperatures for both, the individual energy contributions as well as the thermodynamic temperature used in \( exp(-\Delta G / kT) \)

Deprecated:: Use vrna_exp_params() instead!

Parameters:

temperature – The temperature in degrees Celcius used for (re-)scaling the energy contributions
betaScale – A scaling value that is used as a multiplication factor for the absolute temperature of the system
md – The model details to be used
pf_scale – The scaling factor for the Boltzmann factors

Returns:

A set of precomputed Boltzmann factors

vrna_exp_param_t *get_boltzmann_factor_copy(vrna_exp_param_t *parameters)

#include <ViennaRNA/params/basic.h>

Get a copy of already precomputed Boltzmann factors.

Deprecated:: Use vrna_exp_params_copy() instead!

Parameters:

parameters – The input data structure that shall be copied

Returns:

A copy of the provided Boltzmann factor data set

vrna_exp_param_t *get_scaled_alipf_parameters(unsigned int n_seq)

#include <ViennaRNA/params/basic.h>

Get precomputed Boltzmann factors of the loop type dependent energy contributions (alifold variant)

Deprecated:: Use vrna_exp_params_comparative() instead!

vrna_exp_param_t *get_boltzmann_factors_ali(unsigned int n_seq, double temperature, double betaScale, vrna_md_t md, double pf_scale)

#include <ViennaRNA/params/basic.h>

Get precomputed Boltzmann factors of the loop type dependent energy contributions (alifold variant) with independent thermodynamic temperature.

Deprecated:: Use vrna_exp_params_comparative() instead!

vrna_param_t *scale_parameters(void)

#include <ViennaRNA/params/basic.h>

Get precomputed energy contributions for all the known loop types.

Deprecated:: Use vrna_params() instead!

Note

OpenMP: This function relies on several global model settings variables and thus is not to be considered threadsafe. See get_scaled_parameters() for a completely threadsafe implementation.

Returns:: A set of precomputed energy contributions

vrna_param_t *get_scaled_parameters(double temperature, vrna_md_t md)

#include <ViennaRNA/params/basic.h>

Get precomputed energy contributions for all the known loop types.

Call this function to retrieve precomputed energy contributions, i.e. scaled according to the temperature passed. Furthermore, this function assumes a data structure that contains the model details as well, such that subsequent folding recursions are able to retrieve the correct model settings

Deprecated:: Use vrna_params() instead!

See also

vrna_md_t, set_model_details()

Parameters:

temperature – The temperature in degrees Celcius
md – The model details

Returns:

precomputed energy contributions and model settings

vrna_param_t *copy_parameters(void): #include <ViennaRNA/params/basic.h>

vrna_param_t *set_parameters(vrna_param_t *dest): #include <ViennaRNA/params/basic.h>

vrna_exp_param_t *scale_pf_parameters(void): #include <ViennaRNA/params/basic.h>

vrna_exp_param_t *copy_pf_param(void): #include <ViennaRNA/params/basic.h>

vrna_exp_param_t *set_pf_param(vrna_param_t *dest): #include <ViennaRNA/params/basic.h>

struct vrna_param_s

#include <ViennaRNA/params/basic.h>

The datastructure that contains temperature scaled energy parameters.

Public Members

int id

int stack[NBPAIRS + 1][NBPAIRS + 1]

int hairpin[31]

int bulge[MAXLOOP + 1]

int internal_loop[MAXLOOP + 1]

int mismatchExt[NBPAIRS + 1][5][5]

int mismatchI[NBPAIRS + 1][5][5]

int mismatch1nI[NBPAIRS + 1][5][5]

int mismatch23I[NBPAIRS + 1][5][5]

int mismatchH[NBPAIRS + 1][5][5]

int mismatchM[NBPAIRS + 1][5][5]

int dangle5[NBPAIRS + 1][5]

int dangle3[NBPAIRS + 1][5]

int int11[NBPAIRS + 1][NBPAIRS + 1][5][5]

int int21[NBPAIRS + 1][NBPAIRS + 1][5][5][5]

int int22[NBPAIRS + 1][NBPAIRS + 1][5][5][5][5]

int ninio[5]

double lxc

int MLbase

int MLintern[NBPAIRS + 1]

int MLclosing

int TerminalAU

int DuplexInit

int Tetraloop_E[200]

char Tetraloops[1401]

int Triloop_E[40]

char Triloops[241]

int Hexaloop_E[40]

char Hexaloops[1801]

int TripleC

int MultipleCA

int MultipleCB

int gquad[VRNA_GQUAD_MAX_STACK_SIZE + 1][3 * VRNA_GQUAD_MAX_LINKER_LENGTH + 1]

int gquadLayerMismatch

int gquadLayerMismatchMax

double temperature: Temperature used for loop contribution scaling.

vrna_md_t model_details: Model details to be used in the recursions.

char param_file[256]: The filename the parameters were derived from, or empty string if they represent the default.

int SaltStack

int SaltLoop[MAXLOOP + 2]

double SaltLoopDbl[MAXLOOP + 2]

int SaltMLbase

int SaltMLintern

int SaltMLclosing

int SaltDPXInit

struct vrna_exp_param_s

#include <ViennaRNA/params/basic.h>

The data structure that contains temperature scaled Boltzmann weights of the energy parameters.

Public Members

int id

An identifier for the data structure.

Deprecated:: This attribute will be removed in version 3

double expstack[NBPAIRS + 1][NBPAIRS + 1]

double exphairpin[31]

double expbulge[MAXLOOP + 1]

double expinternal[MAXLOOP + 1]

double expmismatchExt[NBPAIRS + 1][5][5]

double expmismatchI[NBPAIRS + 1][5][5]

double expmismatch23I[NBPAIRS + 1][5][5]

double expmismatch1nI[NBPAIRS + 1][5][5]

double expmismatchH[NBPAIRS + 1][5][5]

double expmismatchM[NBPAIRS + 1][5][5]

double expdangle5[NBPAIRS + 1][5]

double expdangle3[NBPAIRS + 1][5]

double expint11[NBPAIRS + 1][NBPAIRS + 1][5][5]

double expint21[NBPAIRS + 1][NBPAIRS + 1][5][5][5]

double expint22[NBPAIRS + 1][NBPAIRS + 1][5][5][5][5]

double expninio[5][MAXLOOP + 1]

double lxc

double expMLbase

double expMLintern[NBPAIRS + 1]

double expMLclosing

double expTermAU

double expDuplexInit

double exptetra[40]

double exptri[40]

double exphex[40]

char Tetraloops[1401]

double expTriloop[40]

char Triloops[241]

char Hexaloops[1801]

double expTripleC

double expMultipleCA

double expMultipleCB

double expgquad[VRNA_GQUAD_MAX_STACK_SIZE + 1][3 * VRNA_GQUAD_MAX_LINKER_LENGTH + 1]

double expgquadLayerMismatch

int gquadLayerMismatchMax

double kT

double pf_scale: Scaling factor to avoid over-/underflows.

double temperature: Temperature used for loop contribution scaling.

double alpha

Scaling factor for the thermodynamic temperature.

This allows for temperature scaling in Boltzmann factors independently from the energy contributions. The resulting Boltzmann factors are then computed by \( e^{-E/(\alpha \cdot K \cdot T)} \)

vrna_md_t model_details: Model details to be used in the recursions.

char param_file[256]: The filename the parameters were derived from, or empty string if they represent the default.

double expSaltStack

double expSaltLoop[MAXLOOP + 2]

double SaltLoopDbl[MAXLOOP + 2]

int SaltMLbase

int SaltMLintern

int SaltMLclosing

int SaltDPXInit