NEWS | R Documentation |
Changes to functions modelorg2ExPA
and promptSysBiolAlg
,
where closeAllConnections
was called on.exit
of the function.
The two functions now only close the connections that they have opened.
It is now possible to use the functions setSolverParm
and
getSolverParm
in combination with package clpAPI
to set the
number of iterations, to set or get the maximum number of iterations or
maximum number of seconds and to check, if the maximum number of iterations
(or time) was hit. See also setSolverParm-methods
and
getSolverParm-methods
in the manual.
Updated URLs in DESCRIPTION, inst/CITATION and inst/NEWS.Rd and corrected keyword definition for CRAN submission. \
Minor changes removing invalid URLs for CRAN submission. \
Minor code change in addExchReact
for CRAN submission, where
logical operator || encountered and used argument of length more than one.
\
Minor bug fix for function getRedCosts
in optObj_cplexAPI
.
\
Maintainer change to Mayo Roettger <mayo.roettger@hhu.de>. \
Package sybilSBML gets rejected on CRAN, but can be downloaded at https://www.cs.hhu.de/en/research-groups/computational-cell-biology/software.html.
Documentation for react-class
and related methods.
\
addReact
is now a method of the modelorg
class.
Preparation for sybilNLO
.
New reaction object for easy merging and splitting of models. See reactClass
.
Typo in setting methods generated an character vector object
.
\
Minor changes for CRAN submission.
The function genrateWT
caused a problem when setting only solverParm
and no solver.
rmReact
handles now react_attr
and met_attr
corretly, if there is just one column.
The function mod2irrev
crashed when using models with just one gene.
After rmReact
the columns of rxnGeneMat
and
allGenes
-slot did not match.
upgradeModelorg
was not working correctly, when all gprRules
are empty.
modelorg
-class has changed and thus a slot version
was introduced.
the method upgradeModelorg
can convert from earlier instances of this
class. This method is only needed if modelorg
-Objects were loaded
with load
from disk.
gprRules
were stored in an impractical format. This lead to bugs
in rmReact
and further geneDel
.
updated the manual of various files.
doubleGeneDel
messed up the order of deletions when using exlicit gene lists.
corrected url of openCOBRA in the manual.
addReact
crashed for irreversible models.
New Algorithm fbaEasyConstraint
and mtfEasyConstraint
implemented. With these new sysBiolAlg
-Classes it is easier to add
linear constraints to a model. See package vignette or manual for more details.
switch
needs the first parameter explicitly named.
rmReact
error, if resulting model had only one gene left, corrected.
deadEndMetabolite()
could miss deadEndMetabolites if reactions differ in stoichiometry e.g. 1 and 2.
readTSVmod
had a mistake in the documentation about singletons.
changed maintainer
.parseBoolean
now gives a warning, but handles empty gpr in a intuitive way.
findExchangeReact
can now deal with real big matrices (>30.000 columns).
additional example for readProb and writeProb.
modified multiDel
to not use require
doubleFluxDel
had a bug if react1
or react2
contained duplicated entries.
documentation of hasEffect
in optsol_genedel
is now clearer.
fixed a bug in class modelorg
and function addReact
when
the modelorg object is initially empty
for method optimizeProb
of class modelorg
: if arguments
lb
, ub
or obj_coef
are of length one, they are used
for all elements given in react
function geneDel
returns a more illustrative error message if a gene
association is malformed
fixed a bug in optimizer which returns the wrong objective value in rare cases
added boolean parameter fldind
to methods optimizeProb
and applyChanges
which controls the usage of parameter react
function findExchReact
gives a propper warning, if no exchange
reactions can be found and returns NULL
in this case
fixed a bug in function addReact
which accours when more than
one new gene per reaction is added to the model
function checkSolStat
recognizes return code
CPXMIP_OPTIMAL_TOL (102)
: ‘integer optimal, tolerance’
as ‘successfully solved’
added parameter absMAX
to the constructor methods of classes
sysBiolAlg_lmoma
, sysBiolAlg_mtf
and
sysBiolAlg_room
, which is used as a maximum value for upper variable
and contraint bounds
for method optimizeProb
of class modelorg
: if both
arguments gene
and react
are not NULL
, a warning
will be given that react
is ignored
added file ‘CITATION’
added methods singletonMetabolites
and
deadEndMetabolites
to class modelorg
to find dead end
or singleton metabolites in metabolic networks
fixed a bug in function findExchReact
which accours when a
model is in irreversible format
added methods setRowsNames
, getRowsNames
,
setColsNames
and getColsNames
to classes extending
class optObj
added function phpp
for phenotypic phase plane analysis
added class optsol_phpp
added argument rng
to function robAna
fixed a bug in function optimizer
which accours when setting
arguments setToZero
to TRUE
and verboseMode
to
a value larger than two
updated function promptSysBiolAlg
, including now a very basic
example how to set column, row and problem names in problem objects of
class sysBiolAlg
and optObj
fixed a bug in function addReact
which accours if the model
does not contain a GPR association
some adjustments in classes extending class optObj
:
methods loadLPprob
, addRowsToProb
and
changeRowsBnds
works in the same way for each solver (see the
corresponding manual pages for details)
added parameters writeProbToFileName
and pname
to
constructor methods of classes extending class sysBiolAlg
fixed a bug in method changeRowBnds
of class
optObj_lpSolveAPI
suggests on glpkAPI (version >= 1.2.8)
suggests on clpAPI (version >= 1.2.4)
added parameter scaleDist
to the constructor method of class
sysBiolAlg_moma
, which provides a fine tuning method of the
impact of particular reactions on the objective function of the
MOMA algorithm
suggests on glpkAPI (version >= 1.2.7)
suggests on clpAPI (version >= 1.2.3)
variable and constraint names can now be set in objects of class
sysBiolAlg
and optObj
added a new parameter USE_NAMES
added arguments cnames
and rnames
to method
loadLPprob
for classes extending class optObj
added argument rnames
to method addRowsToProb
for
classes extending class optObj
added arguments cnames
and rnames
to method
initialize
of class sysBiolAlg
added arguments useNames
, cnames
and rnames
to
method initialize
of classes sysBiolAlg_fba
,
sysBiolAlg_fv
,sysBiolAlg_lmoma
,sysBiolAlg_moma
,
sysBiolAlg_mtf
and sysBiolAlg_room
.
in method printMetabolite
of class modelorg
:
metabolite id's beginning with a digit or period are prefixed by the
letter "r"
and reaction id's beginning with a digit or period
are prefixed by the letter "x"
function readTSVmod
checks input files for empty fields
added method changeUptake
to class modelorg
to change
media conditions in metabolic networks
fixed a bug in functions .generateWT
and doubleGeneDel
which accours if argument solver was given as variable name and not
as character value directly
added argument termOut
to method initialize
of class
sysBiolAlg
enabling or disabling terminal output of the
mathematical programming software
function checkSolStat
no longer returns an optimal MILP
solution generated by IBM ILOG CPLEX as non-optimal
some refinements for the regular expressions in function
readTSVmod
to identify stoichiometric coefficients and
compartment flags of metabolites
added argument linInd
to function doubleReact
return value of function doubleReact
is now FALSE
if
no identical reactions are found, otherwise a list of id's
corresponding to the reaction id's of identical reactions
default value for argument wtflux
in method initialize
of class sysBiolAlg_moma
is NULL
new parameter to the MOMA algorithm (quadratic version): argument
Qmat
in method initialize
of class
sysBiolAlg_moma
gives the quadratic part of the objective
function
new parameter to the MOMA algorithm (linear version): argument
obj_coefD
in method initialize
of class
sysBiolAlg_lmoma
contains the non-zero part of the objective
function
fixed a bug in function doubleGeneDel
which appears if
argument exLethal
is set to TRUE
removed old file ‘NEWS.Rd’ from top level directory
fixed a bug in the calculation of default objective values in the
mtf
algorithm
method histogram
of class optsol
plots the return value
of mod_obj
, not lp_obj
removed argument MoreArgs
from method optimizeProb
of
class modelorg
slots dels
and fluxdels
of class optsol_fluxdel
and class optsol_genedel
are matrices of class character
class sysBiolAlg
has new methods applyChanges
and
resetChanges
for altering objects of class sysBiolAlg
temporarily
replaced method optimizeProb
of class sysBiolAlg_room
by
methods applyChanges
and resetChanges
constructor method of class sysBiolAlg
has new arguments
retAlgPar
and algPar
used for saving algorithm specific
parameters in objects extending class sysBiolAlg
added slot maxobj
and method changeMaxObj
to class
sysBiolAlg_mtf
fixed a bug in algorithm mtf
: modifications in the model (like
gene knock-outs) were ignored during the calculation of the optimized
objective value
argument algorithm to function optimizer
can now also be
"mtf"
or "fv"
added class netFlux
and constructor function getNetFlux
to order exchange reaction rates according to their sign
function optimizer
returns a list and no longer an object
extending class optsol
.
removed method logOptimizationNE
of class sybilLog
requires R version 2.14.2 or higher
method getFluxDist
of class optsol
get additional
argument opt
to access specific optimization results
removed argument needId
from function checkReactId
removed slot has_id
and method hasId
from
class reactId
removed constructor function reactId
for objects of class
reactId
, use function checkReactId
instead
added function editEnvir
: the environment editor for metabolic
networks
method shrinkMatrix
can get objects of class reactId
removed deprecated functions simpleFBA
and prepProbObj
removed deprecated generics loadProblemData
,
loadProblemDataLM
, loadProblemDataMTF
and loadMatrix
removed deprecated class optsol_simpleFBA
some updates in the vignette
class reactId
is reorganized: it has new slots mod_id
,
mod_key
, react_num
and has_id
and new methods
hasId
and [
, also the implementation of method
show
has changed
new class reactId_Exch
extending class react_id
function findExchReact
now returns an object of class
reactId_Exch
when getting an object of class modelorg
instead of a list
class modelorg
, class reactId
, class optsol
and
class optsolSummary
have new slot mod_key
containing a
character string functioning as a unique key to a model object
fixed a small bug in method printObjFunc
of class
modelorg
fixed a small bug in function fluxVar
which has occurred when
not all reactions given in a model where analyzed
function fluxVar
has new argument exex
in order to
exclude exchange reactions from the analysis
class optsol_fluxVar
has a new method blReact
which does
essentially the same as function blockedReact
function blockedReact
is now deprecated
class optsol
has a new slot obj_func
containing the
objective function of the used model as a character string
method optimizeProb
of class modelorg
returns an object
of class optsol_optimizeProb
as default behaviour (argument
retOptSol
is TRUE
, default)
function geneDel
now returns NULL
instead of NA
if a gene deletion has no effect
default value for arguments gene
, react
, lb
,
ub
and obj_coef
in methods optimizeProb
is
NULL
in optimizeProb
all changes to variable bounds must be set
explicitly, there is no default anymore
added method shrinkMatrix
to class modelorg
for
retrieving subsets of the stoichiometric matrix
changed argument alg
in method initialize
of class
sysBiolAlg
to sbalg
changed argument alg
in function optimizer
to algorithm
slot lp_dir
of class optsol
is now of type factor
solver names "glpk"
, "cplex"
and "clp"
are not
used anymore, use "glpkAPI"
, "cplexAPI"
and
"clpAPI"
instead
added method readProb
to class optObj
removed slot rhs
from class modelorg
, because this
parameter is algorithm dependent and not model dependent
added new class optsolSummary
the function printObjFunc
is now a generic and has a method in
class modelorg
the function checkOptSol
is now a generic and has a method in
class optsol
and class optsolSummary
added methods printMetabolite
and printReaction
to
class modelorg
added method nnzero
to class fluxDistribution
fixed a bug in backupProb methods
fixed a small bug in checkDefaultMethod
moved method mod_obj
from class optsol_fluxdel
to the
base class optsol
added deprecated class optsol_simpleFBA
for compatibility
with other packages depending on sybil
fixed file ‘DESCRIPTION’
some performance improvements in function fluxVar
new function modelorg2ExPA
some very minor improvements in sybilStack (stack [LIFO] and queue [FIFO])
solver names glpk
, cplex
and clp
are deprecated,
use glpkAPI
, cplexAPI
and clpAPI
instead
classes optObj_glpk
, optObj_cplex
and optObj_clp
are renamed to optObj_glpkAPI
, optObj_cplexAPI
and
optObj_clpAPI
fixed a small bug in SYBIL_SETTINGS
appearing if parameters
were set several times
argument solverParm
can be list
or data.frame
argument fmt
to the generic writeProb
is now ff
added support for mixed interger optimization problems with glpkAPI and cplexAPI
added support for quadratic optimization problems with cplexAPI
added methods changeMatrixRow
, loadLPprob
and
loadQobj
to class optObj
cplexAPI must be version 1.2.3 or higher, if used
added function addSolver
in order to make additional solver
packages available
added argument arrowlength
to function readTSVmod
added arguments probType
and loadPackage
to function
checkDefaultMethod
class optsol_doublefluxdel
and class
optsol_doublegenedel
are not available anymore, class
optsol_fluxdel
and class optsol_genedel
are used instead
added new method mod_obj
to class optsol\_fluxdel
, which
returns
c^T v
with c being the vector of
objective coefficients in the model (object of class modelorg
)
and v being the flux
distribution as a result of the optimization. In contrast, method
lp_obj
always returns the value of the objective function
of the used algorithm after optimization.
added functions addAlgorithm
, checkAlgorithm
and
getsybilenv
added a new class sysBiolAlg
including a general constructor
function sysBiolAlg
and subclasses sysBiolAlg_fba
,
sysBiolAlg_fv
, sysBiolAlg_lmoma
, sysBiolAlg_moma
,
sysBiolAlg_mtf
and sysBiolAlg_room
. These classes
contain instances of class optObj
, prepared for specific
algorithms like ROOM or MOMA
support for algorithms MOMA and ROOM in class sysBiolAlg
added method optimizeProb
as a replacement for function
simpleFBA
: unused argument checkIds
is removed, new
arguments gene
and MoreArgs
were added.
added function promptSysBiolAlg
which generates a skeletal
structure for new classes inhereting from sysBiolAlg
functions prepProbObj
and simpleFBA
are deprecated and
will be removed soon, use sysBiolAlg
and optimizeProb
instead
methods loadMatrix
, loadProblemData
,
loadProblemDataLM
and loadProblemDataMTF
are deprecated
and will be removed soon, use loadLPprob
instead
class optsol
makes now use of the package lattice to
plot histograms (use histogram()
instead of hist()
)
updated package vignette
renamed parameter SOLVER_CTRL_PARAM
to SOLVER_CTRL_PARM
renamed argument param
in SYBIL_SETTINGS()
to
parm
file ‘NEWS’ is now in ‘inst/NEWS.Rd’
added argument absCtrl
to the plot
method in
class optsol_robAna
in readTSVmod
, if a model description file is missing, the
basename of the reaction list is used as model id, not the complete
path to the file if fpath is used, or the reaction list is not in
the working directory
removed arguments verboseMode
, loglevel
, logfile
and logfileEnc
from function readTSVmod
; use function
sink
instead
removed arguments loglevel
, logfile
and
logfileEnc
from function optimizer
; use function
sink
instead
added method show
to class sybilError
methods logOptimization
and logOptimizationNE
of class
sybilLog
need argument i
giving a consecutive number
methods logComment
, logOptimization
,
logOptimizationNE
and logOptimizationTH
of class
sybilLog
now also write to stdout
(by cat
) if
verblevel
is > 2
fixed a bug in mod2irrev()
: slots met_de
and
react_de
of class modelorg
(in the reversible model)
are no longer ignored
added function addReact()
to add or change reactions in a
given model
added function addExchReact()
to add exchange reactions for
metabolites to a given model
minor changes in printObjFunc()
SYBIL_SETTINGS(parm) gives an error if parm
is not a valid
parameter name
function findExchReact()
now returns indices of metabolite id's
having an exchange reaction
compatible to clpAPI (version >= 1.2.0)
removed sybilSBML from Suggests field and removed undocumented
function readModel()
added method dim
to class modelorg
renamed method setColsBndsObjCoefs
to
changeColsBndsObjCoefs
updated documentation
depends on glpkAPI (version >= 1.2.1)
needs cplexAPI (version >= 1.2.0)
switched from SparseM to Matrix (version >= 1.0-4), a consequence is a dependency on R (version >= 2.12.0)
completely new implementation of SYBIL_SETTINGS
reorganization of files: one class >= one file
use Collate field in file ‘DESCRIPTION’
LIFO/FIFO (stack/queue) implementation with new functions
stinit()
, stclear()
, stpush()
,
stpop()
,
stunshift()
, stshift()
, stseek()
,
stfirst()
,
stlist()
, stlength()
and stexists()
a few compatibility changes in readTSVmod()
for output of
BiGG database: added arguments
quoteChar
, commentChar
and oneSubSystem
added argument minimalSet
to modelorg2tsv()
added function geneDeletion()
for n gene deletions
fixed a bug in doubleGeneDel()
when allComb = FALSE
added geneDeletion
to multiDel()
added argument checkId
to geneDel()
arguments poCmd
and prCmd
to optimizer
and
simpleFBA
must be of type list
some improvements in sybilLogClass
optimizer()
now makes use of class sybilLog
removed arguments resolve
and resolveParm
from
optimizer()
removed sybil parameters LP_ALT_METHOD
and
SOLVER_ALT_CTRL_PARAM
from SYBIL_SETTINGS()
renamed sybil parameter LP_SOLVER to SOLVER
renamed sybil parameter LP_METHOD to METHOD
moved methods length
and hist
from class
optsol_fluxdel
to class optsol
added slot num_of_prob in class checksol
fixed a bug in class optObj
when setting parameters to
lpSolveAPI
added arguments connect
and colconnect
to the
plot
mathod of class optsol_fluxVar
updated package vignette
fixed a bug in readTSVmod()
when reading closed networks
.createReactionString()
now adds the compartment id to the
metabolite id if it is not already there
if the metabolite id's do not contain the compartment abbreviation in
square brackets, the compartment names are used by
modelorg2tsv()
added arguments exclM
and exclR
to
.singletonMetabolite()
in readTSVmod()
: upper and lower bounds which are larger
than abs(def_bnd)
are set to abs(def_bnd)
added slots met_de
and react_de
to modelorg
added Boolean arguments remMet
, contrMet
,
singletonMet
and deadEndmet
to readTSVmod()
, the
two Boolean arguments remSingleMet
and constrSingleMet
are removed, see help("readTSVmod")
for details
fixed a bug in constructor of class sybilError
depends on cplexAPI (version >= 1.1.7)
checkOptSol()
is compatible with cplexAPI
(version 1.1.7) in combination with IBM ILOG CPLEX (version 12.3)
slots met_single
and react_single
of class
modelorg
are initialized with NA
, not FALSE
some minor changes in readTSVmod()
according to
codetools
fixed a bug in class logFileFH
added slots met_single
and react_single
to class
modelorg
new arguments for readTSVmod()
:
mergeMet
, balanceReact
, remUnusedMetReact
,
remSingleMet
, constrSingleMet
and tol
see
help("readTSVmod")
for details
method logError
of class sybilLog
returns object of
class sybilError
fixed a bug in .printNamedList()
added class sybilLog
added logging in readTSVmod()
, see help("readTSVmod")
for details and new arguments
added NAMESPACE
fixed a minor bug in findExchReact()
fixed a bug in method show
for class SYBIL_SETTINGS
fixed a bug in .validModelorg()
changes in readTSVmod()
:
If a metabolite is used more than once as product or educt of a particular reaction, it is merged:
a + (2) a
is converted to
(3) a
and a warning will be given.
If a metabolite is used first as educt and then as product of a particular reaction, the reaction is balanced:
(2) b + a -> b + c
is converted to
b + a -> c
The stoichiometric matrix S is scanned for metabolites, which are used only once in S. If there are some, a warning will be given.
The stoichiometric matrix S is scanned for reactions, which are not used in S. If there are some, a warning will be given and the corresponding reactions will be removed from the model.
Some improvements in parsing the reaction string: if more than one reaction arrow is found, a warning is given and the reaction will be removed from the model.
support for low level interface glpkAPI (version 1.1.2)
support for low level interface clpAPI (version 1.1.1)
support for low level interface cplexAPI (version 1.1.1)
added new slot in class optObj
: probType
.
The constructor function optObj()
has a new argument
pType
, with lp
as default value (linear programming).
updated documentation
slot subSys
in class modelorg
is now of class
matrix
, so one reaction can belong to more than one subsystem,
for example:
cytosol | mitochondria | external |
|
v_1 | TRUE | FALSE | FALSE |
v_2 | TRUE | TRUE | FALSE |
v_3 | FALSE | TRUE | FALSE |
b_1 | FALSE | FALSE | TRUE |
b_2 | FALSE | FALSE | TRUE |
In slot subSys
, only the column names are named according to
the subsystem. The rows do not have names, the first row corresponds
to the first reaction in slot react_id
, the second row to the
second reaction id, and so forth.
added method getRedCost
to class optObj
.
SBML compatibility is moved to the package sybilSBML, making sybil independend from rsbml. The new package sybilSBML requires a working rsbml installation (available from Bioconductor).
added functions modelorg2tsv()
and readTSVmod()
in order
to handle ‘tsv’ files compatible to the
BiGG database
fixed a bug in optimizer()
when pre/post-processing
should be performed, but no optimization was done.
added argument solverParm
in fluxVar()
, argument
...
is passed to simpleFBA()
now
cleaning of class optObj
:
new classes extending optObj
:
optObj_glpk
, optObj_clp
, optObj_cplex
and optObj_lpSolveAPI
methods setRhsZero
and setColsBndsObjCoefs
are not
usable for initial model creation (this is a feature, not a bug).
Use addRows
and addCols
here.
added stack like functions
added function multiDel()
: multicore
support for
oneGeneDel()
, doubleGeneDel()
, oneFluxDel()
and doubleFluxDel()
fixed a bug in fluxVar()
removed file ‘NAMESPACE’
blockedReact() now works also for irreversible models.
some improvements in readSBMLmod()
depends on rsbml version >= 2.4.0
first public beta release
some things were renamed:
class postAnalysis
is now class ppProc
function postProcessing()
is now function
.ppProcessing()
slot postAna
in class optsol_simpleFBA
is now
slot postProc
argument ppCmd
in function simpleFBA()
is now
argument poCmd
added slot preProc
in class optsol_simpleFBA()
added argument prCmd
in function simpleFBA()
:
same purpose as argument poCmd
, but the commands are
executed immideately before solving the optimization problem.
added slot lethal
to class optsol_doublefluxdel
added method ind2id()
to class optsol_fluxdel
,
optsol_genedel
and optsol_doublegenedel
some minor improvements in doubleGeneDel()
added classes sybilError
and postAnalysis
added function postProcessing()
added slot postAna
in class optsol_simpleFBA
added method sensitivityAnalysis
in class optObj
added argument ppCmd
in simpleFBA()
:
The argument ppCmd
should be a character vector. If it has a
length = 1, it will be treated as a function name, if it has
length > 1, it will be treated as function name (ppCmd[1]
)
plus arguments (ppCmd[-1]
). Basically, ppCmd
is treated as
a command, executed on the problem object (object of class optObj
),
after the problem has been solved. The first element of ppCmd
will
be used as function name, all other elements as arguments to that
function. The command will be joint with ‘,’
and
‘()’
:
ppCmd[1] "(" ppCmd[-1] ")"
and ppCmd[-1]
will be joind with ‘,’
.
The string “LP_PROB”
will be used as a placeholder for the
variable name of the problem object. If “LP_PROB”
is used
in ppCmd
, it will be replaced. The result of the execution of
ppCmd
will be stored in an object of class postAnalysis
in the slot postAna
of class optsol_simpleFBA
.
Some examples:
perform sensitivity analysis with GLPK:
simpleFBA(<model>, ppCmd = "sensitivityAnalysis")
Here, ppCmd
has length = 1,
“sensitivityAnalysis”
will be treated as a function name
and executed on the problem object.
access reduced costs in conjunction with IBM ILOG CPLEX:
simpleFBA(<model>, solver = "cplex", ppCmd = c("getDjCPLEX", "LP_PROB@oobj$env", "LP_PROB@oobj$lp", "0", "react_num(<model>)-1"))
In that example, ppCmd
has length > 1, the first element
is the function name, the others are the arguments. For IBM ILOG CPLEX,
slot oobj
of objects of class optObj
is a list,
containing a pointer to the cplex environment env
and a second
pointer to the problem object lp
. The third and fourth argument
are begin
and end
for the function getDjCPLEX()
.
Remember: in IBM ILOG CPLEX, the first element has index 0!
Instances of class postAnalysis
have two slots:
a single character string containing the executed command
the result of the post processing. If an error occured,
pa
contains the error message and is of class
sybilError
fixed a bug in optimizer()
:
flux distribution was not saved when argument rebuildModel
is
set to TRUE
and the used algorithm is
“linearMOMA”
option resolve = TRUE
in optimizer()
is working again.
setting of parameters to lp solver with SYBIL_SETTINGS
is
working again.
improved function blockedReact()
.
fixed a bug in changeObjCoefs()
fixed a bug in onlyChangeGPR()
fixed a bug in changeGPR()
added function changeGPR()
fixed a small bug in readSBMLmod()
fixed a bug in optimizer()
in conjunction with IBM ILOG CPLEX
and parameter copyModel = TRUE
fixed a bug in optObj_Class
:
loadProblemDataML()
produced an error when
solver = "lpSolveAPI"
and alg = "linearMOMA_COBRA"
method loadProblemDataMTF
in optObj_Class
failed when
IBM ILOG CPLEX is used
fixed a bug in mod2irrev()
: slot rev2irrev
contained
wrong entries.
argument resolve
in optimizer()
is not working in this
version!
completely rewritten function simpleFBA()
; new (changed)
arguments:
(default: NA
) fluxes with changed lower and upper
bounds
(default: NA
) new lower bounds for fluxes in argument
react
(default: NA
) new upper bounds for fluxes in argument
react
(default: FALSE
) Boolean, if set to
TRUE
, after FBA, the sum of all flux rates will be minimized
(default: FALSE
) Boolean, if set to TRUE
,
a minimized distance to a wild type flux distribution will be
generated
(default: NA
) a single flux value (if
argument minTotalFlux
is set to TRUE
) for on optimized
objective function, or a flux vector (if argument minDist
is
set to TRUE
) for an optimized flux distribution.
If wtFluxes
is NA
, FBA will be performed. If arguments
minDist
or minTotalFlux
are TRUE and wtFluxes
is
NA
, the needed values will be calculated by FBA.
added method backupProb()
in class optObj
:
the functions
copyProbGLPK()
/cloneProbCPLEX()
will be used here, which
return a pointer to the new problem object
the functions
saveModelCLP()
/write.lp()
will be used here, which write
the specified problem to file. That file will be deleted after
building the new problem out of those. That may cause a lot IO.
new arguments to optimizer()
[can be passed to
optimizer()
by argument ...
in e.g.
oneGeneDel()
or doubleGeneDel()
]:
(default: FALSE
) Boolean. If set to
TRUE
, the problem object will be recreated for every
optimization. That will increase the running time significantly!
(default: FALSE
) Boolean. If set to TRUE
,
the problem object will be taken from a backup copy. That backup copy
will be created prior to the very first optimization from the modelorg
object.
argument alg
in optimizer()
works now for all solvers
added the used algorithm in the logfile
added dependency on SparseM
added options exLethal
and tol
to doubleGeneDel()
argument alg
in optimizer()
: can be “FBA”
(default), “linearMOMA”
or
“linearMOMA_COBRA”
; works only with GLPK
added method length
for objects of class optsol_fluxdel
added method “[”
for objects of class
optsol_fluxdel
added argument alg
in optimizer()
: can be
“FBA”
(default) or “linearMOMA”
;
works only with glpk
added slot algorithm
and methods algorithm
and
algorithm<-
for objects of class optsol_fluxdel
containing the name of the used algorithm.
added slot TOLERANCE
in SYBIL_SETTINGS
added slot MAXIMUM
in SYBIL_SETTINGS
added slot ALGORITHM
in SYBIL_SETTINGS
added slot OPT_DIRECTION
in SYBIL_SETTINGS
fixed a bug in optObj_Class
, method changeColsBnds
when
using ILOG CPLEX as solver
in optObj_Class
the methods getColsLowBnds
and
getColsUppBnds
now use getLowBndsIdsCPLEX
and
getUppBndsIdsCPLEX
respectively when using ILOG CPLEX as solver
fixed a bug in nonZeroElements()
:
empty rows and columns are now handled correctly when argument
maorder
is set to “column”
or “row”
added support for low level interface sybilCPLEX (version 0.0.2)
minor changes applied to ‘uglyHack.R’ in order to support libsbml version 3.4.1
added setting LP_ALT_METHOD
, used for resolve in
optimizer()
and refineSol()
added class SYBIL_SETTINGS
which contains default settings,
currently for the default lp solver and corresponding method.
The class contains the relevant set/get methods to fit the
content to ones needs.
added method addRowsCols
to class optObj
added support for low level interface lpSolveAPI (version 5.5.0.14)
added support for low level interface sybilCLP (version 0.0.1)
argument resolve
in oneFluxDel()
, oneGeneDel()
,
doubleFluxDel()
and doubleGeneDel()
default: NA
if resolve
is enabled, value should be a valid method
depending on the desired solver (e.g. “exact”
in
case of solver = "glpk"
)
renamed optObj-method setColsBnds
to changeColsBnds
renamed optObj-method setRowsBnds
to changeRowsBnds
renamed optObj-method setObjCoefs
to changeObjCoefs
argument resolve
in optimizer()
and refineSol()
:
no longer Boolean
default: NA
if non optimal solutions should be resolved: set to method resolve should use
argument logfile
in optimizer()
and refineSol()
:
no longer Boolean
default: NA
if a logfile should printed, logfile
will be the filename
argument scaling
in optimizer()
and refineSol()
:
no longer Boolean
default: NA
if the model should be scaled, scaling will be the scaling method
added solver
and method
in ‘logfile’
added argument checkOptSolObj
to function optimizer()
:
print a warning, in cases where solution status != 0
in objects of class optsol (default: FALSE
)
added slots chlb
and chub
to class optsol
:
containing lower and upper bounds of changed flux rates.
replace method: ch?b<-
get method: ch?b
support for low level interface sybilGLPK (version 0.1.0)