ASDmiR: a step-wise method to uncover miRNA regulation related to autism spectrum disorder
MiRNAs (miRNAs) are involved in nervous system developmental, and have potential to cause ASD. However, the miRNA regulation mechanism in ASD is largely unclear. In this work, we present a novel framework, ASDmiR, to identify miRNA-target networks and modules, miRNA sponge networks and modules for uncovering the pathogenesis of ASD, as well as conduct enrichment analysis.
DiffExp_lncR.csv: Differentially expressed lncRNAs.
DiffExp_miR.csv: Differentially expressed miRNAs.
DiffExp_mR.csv: Differentially expressed mRNAs.
DiffExp_miR_lncR.csv: Differentially expressed miRNAs and lncRNAs.
DiffExp_miR_mR.csv: Differentially expressed miRNAs and mRNAs.
miRNA_lncRNA_groundtruth_LncBase_v2.0+NPInter_v4.0.csv: Experimentally validated miRNA-lncRNA interactions from LncBase v2.0 and NPInter v4.0.
miRNA_mRNA_groundtruth_miRTarBase_v8.0+TarBase_v8.0.csv: Experimentally validated miRNA-mRNA interactions from miRTarBase v8.0 and TarBase v8.0.
promise_validated_miR_lncR_mR_1679.el: The format of el about miRNA-target interactions.
ASD.Rdata: ASD expression datasets.
t_lncR_Exp_Autism.csv: LncRNA expression profiles of ASD samples.
t_lncR_Exp_Normal.csv: LncRNA expression profiles of normal samples.
t_miR_Exp_Autism.csv: MiRNA expression profiles of ASD samples.
t_miR_Exp_Normal.csv: MiRNA expression profiles of normal samples.
t_mR_Exp_Autism.csv: MRNA expression profiles of ASD samples.
t_mR_Exp_Normal.csv: MRNA expression profiles of normal samples.
Paste all files into a single folder (set the folder as the directory of R environment), the workflow of ASDmiR is implemented in ASDmiR.R.
For uncovering miRNA regulation related to ASD, we prepare ASD-related miRNA, lncRNA and mRNA expression profiles. Paste the datasets, run script of quick example version (Quick_examples_ASDmiR.R) and source file (zzz.R) into a single folder (set the folder as the directory of R environment).
# Load required R package
library(miRLAB)
library(limma)
# Load utility functions
source("zzz.R")
# Load prepared datasets
raw_Autism <- load("ASD.RData")
# Differentially expressed analysis
DiffExpAna_miR_lncR<-DiffExpAnalysis("t_miR_Exp_Autism.csv", "t_miR_Exp_Normal.csv", "t_lncR_Exp_Autism.csv", "t_lncR_Exp_Normal.csv", topkmiR = 100, topkmR = 300, p.miR = 1, p.mR = 1)
DiffExpAna_miR_mR<-DiffExpAnalysis("t_miR_Exp_Autism.csv", "t_miR_Exp_Normal.csv", "t_mR_Exp_Autism.csv", "t_mR_Exp_Normal.csv", topkmiR = 100, topkmR = 4000, p.miR = 1, p.mR = 1)
# Identification of miRNA-associated regulatory networks by ProMISe
cause = 1:100 #column of 1:35 are miRNAs
effect_lncR = 101:400 #column of 101:400 are lncRNAs
effect_mR = 101:4100 #column of 101:4100 are mRNAs
##predict miRNA targets using ProMISe
DiffExp_promise_miR_lncR <- ProMISe("DiffExp_miR_lncR.csv", cause, effect_lncR)
DiffExp_promise_miR_mR <- ProMISe("DiffExp_miR_mR.csv", cause, effect_mR)
## Select top 200 targets(lncRNAs and miRNAs) of each miRNA
promise_100miR_200lncR = matrix(NA,nrow = 20000, ncol = 3)
for (m in 1:100) {
temp <- as.matrix(bRank(DiffExp_promise_miR_lncR, m, 200, downreg = TRUE))
promise_100miR_200lncR[(200*m-199):(200*m),1:3] = temp[1:200,1:3]
}
colnames(promise_100miR_200lncR) = c("miRNA","gene","Correlation")
promise_100miR_200mR = matrix(NA,nrow = 20000, ncol = 3)
for (m in 1:100) {
temp <- as.matrix(bRank(DiffExp_promise_miR_mR, m, 200, downreg = TRUE))
promise_100miR_200mR[(200*m-199):(200*m),1:3] = temp[1:200,1:3]
}
colnames(promise_100miR_200mR) = c("miRNA","gene","Correlation")
## Validation of miRNA-target interactions predicted by ProMISe method
library(miRLAB)
promise_miR_lncRtop200_Validated <- Validation(promise_100miR_200lncR, datacsv = "miRNA_lncRNA_groundtruth_LncBase_v2.0+NPInter_v4.0.csv")
promise_miR_mRtop200_Validated <- Validation(promise_100miR_200mR, datacsv = "miRNA_mRNA_groundtruth_miRTarBase_v8.0+TarBase_v8.0.csv")
## miRNA-target interactions using ProMISe
promise_validated_miR_lncR_mR_1679 <- rbind(promise_miR_lncRtop200_Validated[[1]],promise_miR_mRtop200_Validated[[1]])
# Identification of miRNA target modules
library(biclique)
library(Rcpp)
bi.format(filename = "promise_validated_miR_lncR_mR_1679.el", filetype = 0)
miR_target_biclique <- bi.clique(filename = "promise_validated_miR_lncR_mR_1679.el", left_least = 3, right_least = 3, filetype = 0)
# Identification of miRNA sponge interaction network by Sensitivity Partial Pearson Correlation (SPPC)
miR_sponge_sppc <- spongeMethod(promise_validated_miR_lncR_mR_1679[,1:2], DEA_miR_mR_lncR, padjustvaluecutoff = 0.05, senscorcutoff = 0.25, method = "sppc")
# Identification of miRNA sponge modules by Markov Cluster Algorithm (MCL) and enrichment analysis.
library(miRspongeR)
miR_sppc_MCL <- netModule(miR_sponge_sppc_0.25[,1:2], method = "MCL",modulesize = 3, save = TRUE)
# Disease and functional enrichment analysis of miRNA target module
library(miRspongeR)
module_DEA <- moduleDEA(miR_sppc_MCL, ont = "DO", OrgDb = "org.Hs.eg.db", padjustvaluecutoff = 0.05, padjustedmethod = "BH")
module_FEA <- moduleFEA(miR_sppc_MCL,ont = "ALL", padjustvaluecutoff = 0.05, padjustedmethod = "BH")