高维状态下肾小管网络的逆向工程matlab代码.zip

function [Matrices] = ORASchafferShrinkageNetworkInference(data,OraSch,LocalFdr,VarInModel) Computing Adjacency Weighted Matrices for networks by using the ORA-Quantile procedure proposed in Pagliarini et al., shrinkage covariace matrix computation plus Local-FDR (Pagliarini et al.), and quantile correlation and partical correlazion plus network deconvolution. Input: data: a n*p matrix, where n is the number of samples and p the number of variables. OraSch: name of the method for the shrinkage estimator of the covariance matrix. If OraSch = 'Sha' the the method is the one proposed in J. Schaefer and K. Strimmer. 2005, OraSch = 'Ora' the method is the one proposed in Chen et al. 2010. LocalFdr: name of the method for network pruning. If LocalFdr = 'Q'the approach is correlation/partial correlation quantile test as described in Pagliarini et al. If LocalFdr = 'L'the approach is the one described in J. Schäfer, K. Strimmer 2005. If LocalFdr = 'N'the approach is correlation/partial correlation quantile test as described in Pagliarini et al. plus network deconvolution developed in Feizi et al. VarInModel: name of the variables/nodes. This input can be empty. Output: a struct array that contains these fields: GGMatrix: the matrix associated with the Graphical Gaussian Model; GGMatrixThre: the pruned GGMatrix; GGMatrixCorrThre: the matrix associated with the output of the ORA-Quantile Algorithm, or with the select pruning method; CorrMatrix: the correlation matrix; OraSch: the input OraSch parameter; LocalFdr: the input LocalFdr parameter; VarInModel: name of the variables/nodes, if this input is not empty. function [Matrices] = ORAShrinkagePartialCorr_V5(data,ShrCov,VarInModel) Computing Adjacency Weighted Matrices for networks by using the ORA-Quantile procedure proposed in Pagliarini et al. and other threshold approached described in the same paper. Input: data: a n*p matrix, where n is the number of samples and p the number of %variables. ShrCov: name of the method for the shrinkage estimator of the covariance matrix. If ShrCov = 'Sha' the the method is the one proposed in J. Schaefer and K. Strimmer. 2005, otherwise the method is the one proposed in Chen et al. 2010. VarInModel: name of the variables/nodes. This input can be empty. Output: a struct array that contains these fields: GGMatrixCorrThre: the matrix associated with the output of the ORA-Quantile Algorithm; CorrMatrixThreLocalFDR: the correlation matrix obtained by using Local-FDR threshold; GGMatrixThreLocalFDR: the matrix associated with the Graphical Gaussian Model obtained by using Local-FDR threshold; GGMatrix: the matrix associated with the Graphical Gaussian Model; CorrMatrix: the correlation matrix; GGMatrixThre: CorrMatrixThre: threshold correlation matrix based on pval t-test; GGMatrixThre: graphical Gaussian model matrix obtained by using the Quantile thresold; GGMatrix_zscore: Graphical Gaussian Model Matrix z-score threshold. Values with absolute value of z-score < 1.96 are set to 0; GGMatrix_2sigma: Graphical Gaussian Model Matrix 2-sigma threshold. Values with absolute value < mu+2*sigma are set to 0, where mu is the mean and sigm is the standard deviation; PvalCorrMatrix: matrix of p-values associated with correlation matrix computed by applying t-test; VarInModel: name of the variables/nodes, if this input is not empty.