Correlations between copy number and mRNA expression
Colon Adenocarcinoma (Primary solid tumor)
02 April 2015  |  analyses__2015_04_02
Maintainer Information
Citation Information
Maintained by John Zhang (MD Anderson Cancer Center)
Cite as Broad Institute TCGA Genome Data Analysis Center (2015): Correlations between copy number and mRNA expression. Broad Institute of MIT and Harvard. doi:10.7908/C14T6HBJ
Overview
Introduction

A TCGA sample is profiled to detect the copy number variations and expressions of genes. This pipeline attempts to correlate copy number and expression data of genes across samples to determine if the copy number variations also result in differential expressions. This report contains the calculated correlation coefficients based on measurements of genomic copy number (log2) values and intensity of the expressions of the corresponding feature across patients. High positive/low negative correlation coefficients indicate that genomic alterations result in differences in the expressions of mRNA the genomic regions transcribe.

Summary

The correlation coefficients in 10, 20, 30, 40, 50, 60, 70, 80, 90 percentiles are -0.047, 0.0157, 0.06792, 0.1243, 0.1905, 0.2636, 0.3381, 0.4178, 0.51226, respectively.

Results
Correlation results

Number of genes and samples used for the calculation are shown in Table 1. Figure 1 shows the distribution of calculated correlation coefficients and quantile-quantile plot of the calculated correlation coefficients against a normal distribution. Table 2 shows the top 20 features ordered by the value of correlation coefficients.

Table 1.  Counts of mRNA and number of samples in copy number and expression data sets and common to both

Category Copy number Expression Common
Sample 450 153 146
Genes 24771 17814 15175

Figure 1.  Summary figures. Left: histogram showing the distribution of the calculated correlations across samples for all Genes. Right: QQ plot of the calculated correlations across samples. The QQ plot is used to plot the quantiles of the calculated correlation coefficients against that derived from a normal distribution. Points deviating from the blue line indicate deviation from normality.

Table 2.  Get Full Table Top 20 features (defined by the feature column) ranked by correlation coefficients

feature r p-value q-value chrom start end geneid
CDC16 0.8451 0 0 8881 13q34 0.343 0.824
CSNK2A1 0.8427 0 0 -6128 20p13 0.715 -0.032
SMARCE1 0.8408 0 0 6605 17q21.2 0.350 0.014
CUL4A 0.8391 0 0 8451 13q34 0.343 0.824
PSMF1 0.8347 0 0 9491 20p13 0.715 -0.032
SAP18 0.8235 0 0 10284 13q12.11 0.343 0.824
CRNKL1 0.8192 0 0 51340 20p11.23 0.715 0.820
ERBB2 0.8131 0 0 2064 17q12 0.350 0.014
SNRPB2 0.8119 0 0 6629 20p12.1 0.715 0.820
PUF60 0.8082 0 0 22827 8q24.3 0.329 0.001
ENTPD6 0.8078 0 0 955 20p11.21 0.715 0.820
YTHDF1 0.8069 0 0 54915 20q13.33 0.708 0.820
TRMT6 0.8067 0 0 51605 20p12.3 0.715 0.849
PCID2 0.8052 0 0 55795 13q34 0.343 0.824
TMTC4 0.7997 0 0 84899 13q32.3 0.343 0.824
RNF6 0.798 0 0 6049 13q12.13 0.343 0.824
MKKS 0.7963 0 0 8195 20p12.2 0.715 0.849
MAP2K4 0.7927 0 0 6416 17p12 -0.330 -0.835
POLR3F 0.7899 0 0 10621 20p11.23 0.715 0.820
DIDO1 0.7898 0 0 11083 20q13.33 0.708 0.820
Methods & Data
Input

Gene level (TCGA Level III) expression data and copy number data of the corresponding loci derived by using the CNTools package of Bioconductor were used for the calculations. Pearson correlation coefficients were calculated for each pair of genes shared by the two data sets across all the samples that were common.

Correlation across sample

Pairwise correlations between the log2 copy numbers and expressions of each gene across samples were calculated using Pearson correlation.

Download Results

In addition to the links below, the full results of the analysis summarized in this report can also be downloaded programmatically using firehose_get, or interactively from either the Broad GDAC website or TCGA Data Coordination Center Portal.