Description Usage Arguments Details Value Author(s) See Also Examples
This function uses FVA under different conditions to find fluxes that linearly correlates to corresponding gene expression.
1 2 3 4 5 6 7 8 9 10 | FECorr (model, nCond, initCond, geneExpressionData=NULL,RuleExpressionData=NULL,
pct_objective=100,
selected_rxns=NULL,
only_identified_rules=FALSE,
minExprFoldChange=0,
lpdir = SYBIL_SETTINGS("OPT_DIRECTION"),
solver = SYBIL_SETTINGS("SOLVER"),
method = SYBIL_SETTINGS("METHOD"),
solverParm=data.frame(CPX_PARAM_EPRHS=1e-6),
verboseMode = 2)
|
model |
An object of class |
nCond |
Number of conditions (FBA problems to be solved) |
geneExpressionData |
a data frame: geneID,Cond_id, ExpressionVal column rows are genes and column j+1 is representing gene expression under condition j |
RuleExpressionData |
rxn_id,cond_id, ExpressionVal |
initCond |
rxn_id,cond_id,lb,ub,objcoef : lower and upper bounds of rxns under different conditions that represents the available nutrients under these conditions |
pct_objective |
Biomass will be garnteed to be at least this value multiplied by max biomass calculated using standard FBA. Values are 0 to 100. |
selected_rxns |
optional parameter used to select a set of reactions not all, Boolean with the same length react_id(model) |
only_identified_rules |
ignore rxns containing genes with unidentified expression |
minExprFoldChange |
can be used to consider only genes with a significant change in expression level (i.e min(expression(gene))*minExprFoldChange*2 must be less than or equal to max(expression(gene))) |
lpdir |
Character value, direction of optimisation. Can be set to |
solver |
Single character string giving the solver package to use. See
|
method |
Single character string giving the method the desired solver has to use.
|
solverParm |
A named data frame or list containing parameters for the specified
solver. Parameters can be set as data frame or list:
|
verboseMode |
An integer value indicating the amount of output to stdout:
0: nothing, 1: status messages, 2: like 1 plus with more details,
3: generates files of the LP problem. |
Main steps 1- Run FVA for all conditions, exclude rxns fixed in all conditions 2- Identify ruleExpression for set of rxns remaining from 1, 3- Fit Expr to FVA range. 4- Run findMDCFlux to find closest genome-scale flux 5- Recalculate correlation: Posterior, iFlux, ruleExpr
returns a list containing slots: geneID,slope,intercept,base_level: OGOR: one gene one rxn iFlux: new fluxes calculated at all the given conditions which considers gene expression data to get a linear fit between gene expression and fluxes. It is a data frame containing the following columns rxn_id: reactionId in model,cond_id:,lb,ub,objCoef,xpc_flux,fva_min,fva_max,RuleExprVal,iflux
Abdelmoneim Amer Desouki
modelorg
,
optimizeProb
,
gene2Rule
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 | ## Not run:
library(sybil)
data(iAF1260)
model= iAF1260
# trivial test 1, linear levels for an unbounded rxn
ncnd=3
rxn="R_ADK1"
nc=length(react_id(model))
initCond=cbind(rxn_id=react_id(model),cond_id=rep(1,nc),
lb=lowbnd(model),ub=uppbnd(model),obj=obj_coef(model))
initCond=rbind(initCond,cbind(rxn_id=react_id(model),
cond_id=rep(2,nc),lb=lowbnd(model),ub=uppbnd(model),
obj=obj_coef(model)))
initCond=rbind(initCond,cbind(rxn_id=react_id(model),
cond_id=rep(3,nc),lb=lowbnd(model),ub=uppbnd(model),
obj=obj_coef(model)))
cnds=(1:3)
gprExp=cbind(rxn_id=rxn,cond_id=1,expr_val=2)
gprExp=rbind(gprExp,cbind(rxn_id=rxn,cond_id=2,expr_val=4))
gprExp=rbind(gprExp,cbind(rxn_id=rxn,cond_id=3,expr_val=6))
fcflx=FECorr(model,nCond=ncnd,initCond=initCond,
RuleExpressionData=gprExp,selected_rxns=(react_id(model)==rxn),
verboseMode=4);
fcflx[[2]][fcflx[[2]][,2]==rxn,]
cor(as.numeric(fcflx[[2]][fcflx[[2]][,2]==rxn,"expr_val"]),
as.numeric(fcflx[[2]][fcflx[[2]][,2]==rxn,"iflx"]))
## End(Not run)
|
Add the following code to your website.
For more information on customizing the embed code, read Embedding Snippets.