An integrative analysis tool to screen for Proteogenomic Functional traits perturbed by DNA copy number alterations (CNA) and DNA methylation
The goal of iProFun is to
-
characterize functional consequences of DNA copy number and methylation alterations in tumors
-
facilitate screening for cancer drivers contributing to tumor initiation and progression, since CNAs and DNA methylations that preserve functional consequences are more likely to be cancer drivers.
This package implement iProFun using 2 main functions. The primary function is (surprise!) iProFun, which first fit multiple linear regression to obtain summary statistics and then use the statistics to detect joint associations of DNA alternations with multi-omic traits. False discovery rate assessment is supported by iProFun_permutate function. A full description of the method can be found in our paper.
You can install the latest version directly from GitHub with devtools:
install.packages("devtools")
devtools::install_github("songxiaoyu/iProFun")Below is an example of iProFun Integrative analysis pipeline.
The sample data: cna, methy, rna, protein, phospho, rna_pc_1_3, protein_pc_1_3, phospho_pc_1_3 are included in the package. A full description of the sample dataset can be found in the help page of the data.
library(iProFun)
data(cna)
?cnaTo implement iProFun, we can use iProFun
iprofun_result <- iProFun(ylist = list(rna, protein, phospho), xlist = list(cna, methy), covariates = list(rna_pc_1_3,
protein_pc_1_3, phospho_pc_1_3), pi = rep(0.05, 3))## [1] "No IDs are specified. The first column of each data type in ylist and xlist are used."
## [1] "A total of 676 genes/proteins/peptides/etc. are considered."
## [1] "A total of 525 samples for Y1 data type."
## [2] "A total of 171 samples for Y2 data type."
## [3] "A total of 68 samples for Y3 data type."
## [1] "Obtaining regression summaries for data type 1"
## [1] "Obtaining regression summaries for data type 2"
## [1] "Obtaining regression summaries for data type 3"
## [1] "Run iProFun"
## [1] "Completed"
The result of the function contains a list. The length of the output list equals to the number of element in the xlist. In each element of the list, there is a sublist containing the results associated with the data in xlist.
The following componentshows the marginal posterior probability for each pattern associated with the first element in xlist.
iprofun_result$`iProFun output for xlist 1`$colocProb## [1] 1.620700e-02 7.908244e-02 1.803997e-10 3.075494e-07 4.505220e-01
## [6] 1.341961e-03 3.813558e-13 4.528463e-01
We can also show the posterior probability for each pattern associated with the first element in xlist.
head(cbind(iprofun_result$`iProFun output for xlist 1`$yName_J, iprofun_result$`iProFun output for xlist 1`$PostProb))## phospho_ID
## 1 "AAAS" "AAAS.NP_001166937.1:s462" "3.9237564406578e-47"
## 2 "ABCC1" "ABCC1.NP_004987.2:s930" "2.56137582110385e-24"
## 4 "ABI1" "ABI1.NP_001012768.1:s222" "1.40677992841573e-41"
## 5 "ABI2" "ABI2.NP_001269854.1:s183" "1.87012825266245e-64"
## 9 "ABLIM1" "ABLIM1.NP_001309811.1:s420" "4.94291355528985e-31"
## 16 "ACLY" "ACLY.NP_001087.2:s481" "1.76491191405199e-26"
##
## 1 "0.0123749091873888" "8.96383959025961e-54" "1.54463510952455e-51"
## 2 "0.000120595712748723" "8.71286778050195e-29" "4.49085550402044e-29"
## 4 "0.00462301449129315" "1.40567545207946e-47" "1.64341316443298e-46"
## 5 "4.50847967166922e-06" "6.08461639892011e-68" "2.0230991904436e-68"
## 9 "3.20999177143844e-07" "7.66963265890669e-35" "2.0376595190263e-33"
## 16 "0.00733202484142715" "6.64945949316372e-33" "7.31464611758034e-31"
##
## 1 "0.449574541972872" "0.000326191725840336" "3.64364112491387e-56"
## 2 "0.652360027007038" "1.41577651564761e-06" "1.57737356460954e-31"
## 4 "0.734601494376669" "3.61620391871919e-05" "1.69559964669517e-50"
## 5 "0.233270372819702" "3.26575018606226e-07" "6.7966772468376e-70"
## 9 "0.00792069194509253" "8.86052756345061e-07" "3.26468154769303e-35"
## 16 "0.439293617050499" "0.000203470415596139" "2.84560285735056e-35"
##
## 1 "0.537724357113899"
## 2 "0.347517961503697"
## 4 "0.26073932909285"
## 5 "0.766724792125608"
## 9 "0.992078101002974"
## 16 "0.553170887692478"
The following shows the beta coefficients for each gene-level multiple linear regression.
head(cbind(iprofun_result$`iProFun output for xlist 1`$yName_J, iprofun_result$`iProFun output for xlist 1`$betas_J))## phospho_ID
## 1 "AAAS" "AAAS.NP_001166937.1:s462" "2.15294856564555"
## 2 "ABCC1" "ABCC1.NP_004987.2:s930" "1.31481344483779"
## 4 "ABI1" "ABI1.NP_001012768.1:s222" "1.70976139319942"
## 5 "ABI2" "ABI2.NP_001269854.1:s183" "2.30305705519527"
## 9 "ABLIM1" "ABLIM1.NP_001309811.1:s420" "1.18040186037167"
## 16 "ACLY" "ACLY.NP_001087.2:s481" "1.14383323720261"
##
## 1 "0.786244511025172" "0.764626299444886"
## 2 "1.04663228802151" "0.491584008481416"
## 4 "0.816841442720532" "-0.2947058215859"
## 5 "1.45245476052363" "1.07690206117779"
## 9 "1.01209731356699" "1.59586460265822"
## 16 "0.518544940554173" "0.704512877526666"
To control false discovery rate for the result, we can use iProFun_permutate. To save time, I will set number of permutation to 1. In our paper, we permutate 100 times to control the empirical FDR.
iprofun_permutate_result <- iProFun_permutate(ylist = list(rna, protein, phospho),
xlist = list(cna, methy), covariates = list(rna_pc_1_3, protein_pc_1_3, phospho_pc_1_3),
pi = rep(0.05, 3), permutate_number = 1, fdr = 0.1, PostCut = 0.75, filter <- c(1,0),
grids = c(seq(0.75, 0.99, 0.01), seq(0.991, 0.999, 0.001), seq(0.9991, 0.9999, 0.0001)),
seed=123)## [1] "Finish permutation 1"
## Warning in min(which(f2 < fdr)): no non-missing arguments to min; returning
## Inf
The result of the function contains a list with the following components. The first component shows the cutoff values for each group based on permutation.
iprofun_permutate_result$fdr_cutPob## [[1]]
## [1] 0.75 0.79 0.75
##
## [[2]]
## [1] 0.75 0.75 NA
The following components show the genes names that are identified by iProFun with the FDR crteria we set in each group.
head(iprofun_permutate_result$Gene_fdr[[1]])## Y1 Y2 Y3
## [1,] "AAAS" "1" "1" "0"
## [2,] "ABCC1" "1" "1" "0"
## [3,] "ABI1" "0" "0" "0"
## [4,] "ABI2" "1" "1" "1"
## [5,] "ABLIM1" "1" "1" "1"
## [6,] "ACLY" "1" "1" "0"
If you find small bugs, larger issues, or have suggestions, please file them using the issue tracker or email the maintainer at Jiayi.Ji@mountsinai.org. Contributions (via pull requests or otherwise) are welcome.