| Title: | S-system parameter estimation method |
|---|---|
| Description: | This package can optimize the parameter in S-system models given time series data |
| Authors: | Xinyi YANG Developer, Jennifer E. DENT Developer and Christine NARDINI Supervisor |
| Maintainer: | Xinyi YANG <[email protected]> |
| License: | GPL-2 |
| Version: | 1.45.0 |
| Built: | 2024-07-03 06:02:32 UTC |
| Source: | https://github.com/bioc/SPEM |
The function in this package allows for the computation of parameters in the n-gene S-system from time series data.
| Package: | SPEM |
| Type: | Package |
| Version: | 1.0 |
| Date: | 2013-01-13 |
| License: | GPL-2 |
| LazyLoad: | yes |
Yang, X-Y, Dent, J. E. and Nardini, C.
Maintainer: Xinyi Yang [email protected]
This function calculates parameters for a single row in the expression data. If a large-size dataset will be calculated, this function is recommended.
## S4 method for signature 'ExpressionSet' row_optimize(TS_eSet,S,beta, sparsity = 0.2, lbH = -3, ubH = 3, lbB = 0, ubB = 10)## S4 method for signature 'ExpressionSet' row_optimize(TS_eSet,S,beta, sparsity = 0.2, lbH = -3, ubH = 3, lbB = 0, ubB = 10)
TS_eSet |
Time series data in ExpressionSet class assayData: Matrix with n metabolite in row and m time points in column. phenoData: Dataframe includes label "time", which represents the time points. |
S |
Slope of the row you want to calculated. You can either input a vector with length equal to the rows of assayData of TS_eSet, or use s_diff function in this package to calculate it. |
beta |
Initial beta. |
sparsity |
A threshold used to control the sparsity of reconstructed matrix. Values whose absolute value smaller than sparsity will be set to zero. |
lbH |
Lower boundary value of h. |
ubH |
Upper boundary value of h. |
lbB |
Lower boundary value of beta. |
ubB |
Upper boundary value of beta. |
In this SPEM package, we aim to reconstruct gene networks from time-series expression data using the S-system model. The input dataset should be as an ExpressionSet data container, describing, in assayData, expression data for n genes (rows) and m time points (columns), along with a vector of length m, which records the exact values of time points, thus showing the sample intervals in phenoData. SPEM will calculate the parameters alpha, g, beta and h of the S-system function set that best fits the dataset.
In this function, user can calculate one row at a time. This function offers a parallel calculation option for users.
This function return a vector bind with c(alpha, $g_i$, beta, $h_i$, Initial Beta, error).
signature(TS_eSet = "ExpressionSet")This method is created for the function row_optimize.
Yang, X-Y, Dent, Jennifer E. and Nardini, C.
#########Load the SOS pathway data ####### data(sos) #########Set Slope and Initial Beta ####### Slope<- s_diff(sos) S<- Slope[1,] #S is the slope of the row you want to calculate. You can either input a vector yourself. beta<- runif(n=1,min=1,max=10) #########Set parameters ####### sparsity<- 0.2 lbH<- -3 ubH<- 3 lbB<- 0 ubB<- 10 #########Calculate results ####### result_r<-row_optimize(sos,S,beta,sparsity,lbH,ubH,lbB,ubB)#########Load the SOS pathway data ####### data(sos) #########Set Slope and Initial Beta ####### Slope<- s_diff(sos) S<- Slope[1,] #S is the slope of the row you want to calculate. You can either input a vector yourself. beta<- runif(n=1,min=1,max=10) #########Set parameters ####### sparsity<- 0.2 lbH<- -3 ubH<- 3 lbB<- 0 ubB<- 10 #########Calculate results ####### result_r<-row_optimize(sos,S,beta,sparsity,lbH,ubH,lbB,ubB)
This function allows users calculate slopes from time points and time series data.
## S4 method for signature 'ExpressionSet' s_diff(TS_eSet)## S4 method for signature 'ExpressionSet' s_diff(TS_eSet)
TS_eSet |
Time series data in ExpressionSet class. assayData: Matrix with n metabolite in row and m time points in column. phenoData: phenoData type. The sample data.frame should include the label "time", which represents the values of time points. |
This function directly return a slope matrix.
signature(TS_eSet = "ExpressionSet")This method is created for function s_diff.
Yang, X-Y, Dent, Jennifer E. and Nardini, C.
#########Load the SOS pathway data ####### data(sos) #########Calculate results ####### Slope<-s_diff(sos)#########Load the SOS pathway data ####### data(sos) #########Calculate results ####### Slope<-s_diff(sos)
In this package we offer the SOS data obtained from Uri Alon's lab (http://www.weizmann.ac.il/mcb/UriAlon/). SOS response is a general DNA repair system in bacteria which allows survival after DNA damage. This SOS dataset is taken from real experiment expression data in Escherichia coli. It contains 8 genes under Experiment 3 (UV light intensities, 4:20 Jm^(-2)).
data(sos)data(sos)
sos.data is time series gene expression value data in ExpressionSet Class. assayData: Matrix with expression values of 8 genes in SOS pathway of Escherichia coli. These expression levels are observed at 50 time points. phenoData: Sample data.frame includes label "time", which represents the value of time points.
M. Ronen, R. Rosenberg, B. I. Shraiman, and U. Alon. Assigning numbers to the arrows: parameterizing a gene regulation network by using accurate expression kinetics. Proceedings of the National Academy of Sciences of the United States of America, Aug. 2002. PMID: 12145321.
data(sos)data(sos)
This function calculates parameters of S-system from entire time series matrix.
## S4 method for signature 'ExpressionSet' SPEM(TS_eSet, n = 3, sparsity = 0.2, lbH = -3, ubH = 3, lbB = 0, ubB = 10)## S4 method for signature 'ExpressionSet' SPEM(TS_eSet, n = 3, sparsity = 0.2, lbH = -3, ubH = 3, lbB = 0, ubB = 10)
TS_eSet |
Time series data in ExpressionSet class. assayData: Matrix with n metabolite in row and m time points in column. phenoData: phenoData type. The sample data.frame should include the label "time", which represents the values of time points. |
n |
Positive integer, SPEM will guess initial beta n times. |
sparsity |
A positive number. In order to force the interaction matrix to be sparse, interactions with absolute value smaller than "sparsity" will be set to zero. |
lbH |
Lower boundary value of h. |
ubH |
Upper boundary value of h. |
lbB |
Lower boundary value of beta. |
ubB |
Upper boundary value of beta. |
In this SPEM package, we aim to reconstruct gene networks from time-series expression data using the S-system model. The input dataset should be as an ExpressionSet data container, describing, in assayData, expression data for n genes (rows) and m time points (columns), along with a vector of length m, which records the exact values of time points, thus showing the sample intervals in phenoData. SPEM will calculate the parameters alpha, G, beta and H of the S-system function set that best fits the dataset.
alpha, G, beta, H
|
Parameters of the reconstructed S-system. |
IniBeta |
Guess of the IniBeta value (Picked randomly by SPEM itself). |
error |
Regression error. |
signature(TS_eSet = "ExpressionSet")This method is created for function SPEM.
Yang, X-Y., Dent, Jennifer E. and Nardini, C.
#########Generate Toy Model ####### ######### # If you want to calculate SOS dataset in this package, please read our vignette### #Real dataset takes a long time to calculate. You may want to try function 'row_optimize' to compute it in parallel### toy_expression_data<-matrix(data=abs(rnorm(12)),nrow=3,ncol=4, dimnames=list(paste("G",c(1:3),sep=''), paste("tp",c(0,2,4,6),sep="_"))) toy_timepoints_data<-data.frame(index=c(1:4), label=paste("tp",c(0,2,4,6),sep='_'), time=c(0,2,4,6),row.names=paste("tp",c(0,2,4,6),sep='_')) toy_varMetadata<-data.frame(labelDescription=c("Index number","Label Detail", "Time points values"),row.names=c("index","label","time")) toy_phenoData<-new("AnnotatedDataFrame", data=toy_timepoints_data,varMetadata=toy_varMetadata) toy_ExpressionSet<-new("ExpressionSet", exprs=toy_expression_data,phenoData=toy_phenoData) #########Set parameters ####### n<- 1 sparsity<- 0.2 lbH<- -3 ubH<- 3 lbB<- 0 ubB<- 10 #########Calculate results ####### result<-SPEM(toy_ExpressionSet,n,sparsity,lbH,ubH,lbB,ubB)#########Generate Toy Model ####### ######### # If you want to calculate SOS dataset in this package, please read our vignette### #Real dataset takes a long time to calculate. You may want to try function 'row_optimize' to compute it in parallel### toy_expression_data<-matrix(data=abs(rnorm(12)),nrow=3,ncol=4, dimnames=list(paste("G",c(1:3),sep=''), paste("tp",c(0,2,4,6),sep="_"))) toy_timepoints_data<-data.frame(index=c(1:4), label=paste("tp",c(0,2,4,6),sep='_'), time=c(0,2,4,6),row.names=paste("tp",c(0,2,4,6),sep='_')) toy_varMetadata<-data.frame(labelDescription=c("Index number","Label Detail", "Time points values"),row.names=c("index","label","time")) toy_phenoData<-new("AnnotatedDataFrame", data=toy_timepoints_data,varMetadata=toy_varMetadata) toy_ExpressionSet<-new("ExpressionSet", exprs=toy_expression_data,phenoData=toy_phenoData) #########Set parameters ####### n<- 1 sparsity<- 0.2 lbH<- -3 ubH<- 3 lbB<- 0 ubB<- 10 #########Calculate results ####### result<-SPEM(toy_ExpressionSet,n,sparsity,lbH,ubH,lbB,ubB)