Added Kosts Method for Matlab

Matlab/KostsMethod.m

@@ -0,0 +1,69 @@
+% Copyright 2015, Institute for Systems Biology.
+% 
+% Licensed under the Apache License, Version 2.0 (the "License");
+% you may not use this file except in compliance with the License.
+% You may obtain a copy of the License at
+% 
+% http://www.apache.org/licenses/LICENSE-2.0
+% 
+% Unless required by applicable law or agreed to in writing, software
+% distributed under the License is distributed on an "AS IS" BASIS,
+% WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+% See the License for the specific language governing permissions and
+% limitations under the License.
+% 
+% Author: William Poole
+% Ported to Matlab: Theo Knijnenburg
+% Email: [email protected] / [email protected]
+% Created: February 2016
+
+function [Pkost,Pfisher,Ckost,DFkost] = KostsMethod(data_matrix, p_values)
+
+%license, author, date
+
+%% Inputes and outputs
+
+% Input:  An m x n data matrix with each of m rows representing a variable and each of n columns representing a sample. Should be of type numpy.array
+%         A vector of m P-values to combine. May be a list or of type numpy.array.
+% Output: A combined P-value.
+%         If extra_info == True: also returns the p-value from Fisher's method, the scale factor c, and the new degrees of freedom from Brown's Method
+
+%% Compute P-values
+covar_matrix = CalculateCovariances(data_matrix);
+[Pkost,Pfisher,Ckost,DFkost] = CombinePValues(covar_matrix, p_values);
+
+%% Functions
+
+%Calculate Covariances
+    function covar_matrix = CalculateCovariances(data_matrix);
+        [m,n] = size(data_matrix);
+        cor = corr(data_matrix');
+        %Kost's polynomial fit
+        a1 = 3.263;
+        a2 = 0.710; 
+        a3 = 0.027;
+        covar_matrix = a1.*cor+a2.*cor.^2+a3.*cor.^3;
+    end
+
+% Combining P-values
+    function [p_kost,p_fisher,c,df_kost] = CombinePValues(covar_matrix, p_values);
+
+        m = size(covar_matrix,1);
+        df_fisher = 2.0*m;
+        Expected = 2.0*m;
+        cov_sum = sum(sum(covar_matrix))-sum(diag(covar_matrix));
+        Var = 4.0*m+cov_sum;
+        c = Var/(2*Expected);
+        df_kost = 2*(Expected^2)/Var;
+        if df_kost > df_fisher;
+            df_kost = df_fisher;
+            c = 1;
+        end
+
+        x = 2*sum(-log(p_values));
+        p_kost = chi2cdf(1.0*x/c,df_kost,'upper');
+        p_fisher = chi2cdf(1.0*x,df_fisher,'upper');
+    end
+
+end
+

Matlab/WorkFlow.m

@@ -16,6 +16,7 @@
 % Ported to Matlab: Theo Knijnenburg
 % Email: [email protected] / [email protected]
 % Created: June 2015
+% Updated: February 2016
 
 %% Initialize
 Init
@@ -28,7 +29,7 @@
 %% Pearson correlation
 [R,P] = corr(D(1,:)',D(2:end,:)');
 
-%% Emperical Browns Methods
+%% Emperical Browns Method
 data_matrix = D(2:end,:);
 p_values = P;
 [Pbrown,Pfisher,Cbrown,DFbrown] = EmpiricalBrownsMethod(data_matrix, p_values)
@@ -54,6 +55,30 @@
 % 
 %     8.1367
 
+%% Kosts Method
+[Pkost,Pfisher,Ckost,DFkost] = KostsMethod(data_matrix, p_values)
+
+%% Should give...
+
+% Pkost =
+% 
+%     0.7018
+% 
+% 
+% Pfisher =
+% 
+%     0.8614
+% 
+% 
+% Ckost =
+% 
+%     2.8144
+% 
+% 
+% DFkost =
+% 
+%     7.1063
+
 
 %% Load TCGA dataset
 % parseTCGAdata
@@ -70,11 +95,12 @@
 
     display(Pathways{p,1});
     [Pbrown,Pfisher,Cbrown,DFbrown] = EmpiricalBrownsMethod(data_matrix, p_values)
-
+    [Pkost,Pfisher,Ckost,DFkost] = KostsMethod(data_matrix, p_values)
 end
 
 
 %% Should give...
+
 % FOXA1 TRANSCRIPTION FACTOR NETWORK
 % 
 % Pbrown =
@@ -96,6 +122,26 @@
 % 
 %    21.3285
 % 
+% 
+% Pkost =
+% 
+%    3.1577e-57
+% 
+% 
+% Pfisher =
+% 
+%   4.0434e-139
+% 
+% 
+% Ckost =
+% 
+%     2.5009
+% 
+% 
+% DFkost =
+% 
+%    23.1920
+% 
 % GLYPICAN 3 NETWORK
 % 
 % Pbrown =
@@ -117,6 +163,26 @@
 % 
 %    10.7884
 % 
+% 
+% Pkost =
+% 
+%    7.5708e-07
+% 
+% 
+% Pfisher =
+% 
+%    1.4387e-08
+% 
+% 
+% Ckost =
+% 
+%     1.3490
+% 
+% 
+% DFkost =
+% 
+%    10.3777
+% 
 % SUMOYLATION BY RANBP2 REGULATES TRANSCRIPTIONAL REPRESSION
 % 
 % Pbrown =
@@ -137,7 +203,26 @@
 % DFbrown =
 % 
 %    18.3869
-
+% 
+% 
+% Pkost =
+% 
+%    2.0949e-39
+% 
+% 
+% Pfisher =
+% 
+%    6.4438e-45
+% 
+% 
+% Ckost =
+% 
+%     1.1494
+% 
+% 
+% DFkost =
+% 
+%    17.4011
 
 
 

R/.Rhistory

@@ -0,0 +1,43 @@
+source("EmpericalBrownsMethod/R/ebm.R")
+options(digits=16)
+# Test with artificical data
+print("****************************************************************")
+print("RANDOM DATA")
+randData <- read.table("../Data/RandomData.tsv", sep="\t", header=F, stringsAsFactors=F)
+a <- as.numeric(randData[1,-1])
+rd <- randData[-1,-1]
+pvals <- sapply(1:10, function(i) cor.test(a, as.numeric(rd[i,]))$p.value)
+print(empiricalBrownsMethod(data_matrix=rd, p_values=pvals, extra_info=T))
+print("****************************************************************")
+source("EmpericalBrownsMethod/R/ebm.R")
+source("/EmpericalBrownsMethod/R/ebm.R")
+source("EmpiricalBrownsMethod/R/ebm.R")
+options(digits=16)
+print("****************************************************************")
+print("RANDOM DATA")
+randData <- read.table("../Data/RandomData.tsv", sep="\t", header=F, stringsAsFactors=F)
+a <- as.numeric(randData[1,-1])
+rd <- randData[-1,-1]
+pvals <- sapply(1:10, function(i) cor.test(a, as.numeric(rd[i,]))$p.value)
+print(empiricalBrownsMethod(data_matrix=rd, p_values=pvals, extra_info=T))
+print("****************************************************************")
+print(KostsMethod(data_matrix=rd, p_values=pvals, extra_info=T))
+print("****************************************************************")
+print(kostsMethod(data_matrix=rd, p_values=pvals, extra_info=T))
+print("****************************************************************")
+print(kostsMethod(data_matrix=as.matrix(rd), p_values=pvals, extra_info=T))
+print("****************************************************************")
+pathways <- read.table("../Data/pathways.tsv", sep="\t", header=T, stringsAsFactors=F)
+allPvals <- read.table("../Data/CDH4_Pvalues.tsv", sep="\t", stringsAsFactors=F, header=T )
+dat <- read.table("../Data/ReducedFeatureMatrix.tsv", sep="\t", stringsAsFactors=F, header=F)
+allPvals <- (unique(allPvals))
+print("Glypican 3 Network")
+glypGenes <- pathways$gene[pathways$pathway == "GLYPICAN 3 NETWORK"]
+glypPvals <- allPvals$pvalue.with.CHD4[allPvals$gene %in% glypGenes]
+glypDat <- dat[dat$V1 %in% glypGenes, 2:ncol(dat)]
+print(empiricalBrownsMethod(data_matrix=glypDat, p_values=glypPvals, extra_info=T))
+print("****************************************************************")
+print(kostsMethod(data_matrix=glypDat, p_values=glypPvals, extra_info=T))
+print("****************************************************************")
+print(kostsMethod(data_matrix=as.matrix(glypDat), p_values=glypPvals, extra_info=T))
+print("****************************************************************")

R/WorkFlow.R

@@ -16,8 +16,9 @@
 # Ported to R: David L Gibbs
 # Email: [email protected] / [email protected] / [email protected]
 # Created: June 2015
+# Updated: February 2016
 
-source("EmpericalBrownsMethod/R/ebm.R")
+source("EmpiricalBrownsMethod/R/ebm.R")
 options(digits=16)
 
 # Test with artificical data
@@ -44,6 +45,23 @@ print("****************************************************************")
 #$DF_Brown
 #[1] 8.136664603851868
 
+print(kostsMethod(data_matrix=as.matrix(rd), p_values=pvals, extra_info=T))
+print("****************************************************************")
+
+#Should give:
+# $P_test
+# [1] 0.701752883272515
+# 
+# $P_Fisher
+# [1] 0.8613842570343421
+# 
+# $Scale_Factor_C
+# [1] 2.814405567447344
+# 
+# $DF
+# [1] 7.106296345959808
+
+
 # Test with cancer data #
 
 pathways <- read.table("../Data/pathways.tsv", sep="\t", header=T, stringsAsFactors=F)
@@ -71,6 +89,22 @@ print("****************************************************************")
 #$DF_Brown
 #[1] 10.78837806737645
 
+print(kostsMethod(data_matrix=as.matrix(glypDat), p_values=glypPvals, extra_info=T))
+print("****************************************************************")
+
+#Should give:
+# $P_test
+# [1] 7.570776008807138e-07
+# 
+# $P_Fisher
+# [1] 1.438732140605804e-08
+# 
+# $Scale_Factor_C
+# [1] 1.349048766471012
+# 
+# $DF
+# [1] 10.3776826664485
+
 print("SUMO Reg")
 sumoGenes <- pathways$gene[pathways$pathway == "SUMOYLATION BY RANBP2 REGULATES TRANSCRIPTIONAL REPRESSION"]
 sumoPvals <- allPvals$pvalue.with.CHD4[allPvals$gene %in% sumoGenes]

gitsteps.txt

@@ -0,0 +1,14 @@
+(on remus)
+
+git init
+git pull https://github.com/IlyaLab/CombiningDependentPvaluesUsingEBM
+
+git status
+git add *
+git status
+git commit -m 'Initial commit'
+git push https://github.com/IlyaLab/CombiningDependentPvaluesUsingEBM
+
+
+
+