# load the CORNA package

library(CORNA)

# read mirbase targets data for Mus musculus

targets <- miRBase2df.fun(file="arch.v5.txt.mus_musculus.zip")

# obtain a link from transcript to gene from biomart

tran2gene <- BioMart2df.fun(
                biomart="ensembl",
                dataset="mmusculus_gene_ensembl",
                col.old=c("ensembl_transcript_id",
                          "ensembl_gene_id"),
                col.new=c("tran", "gene"))

# link microRNAs to genes instead of transcripts

mir2gene <- corna.map.fun(targets,
                          tran2gene,
                          "gene",
                          "mir")

# get those genes associated with mmu-mir-155

gvec <- corna.map.fun(mir2gene,
                      "mmu-mir-155",
                      "mir",
                      "gene")

# read pathway information from KEGG for mmu

gene2path <- KEGG2df.fun(org="mmu")

# sample is only those genes associated with mmu-mir-155 that have pathway data

gvec <- intersect(gvec, unique(gene2path$gene))

# create data frame of pathway ids and names

path2name <- unique(gene2path[c("path", "name")])

# perform the test

test <- corna.test.fun(
            gvec,
            unique(gene2path$gene),
            gene2path,
            hypergeometric=T,
            fisher=T,
            chi.square=T,
            #fisher.alternative="greater",
            min.pop=10,
            sort="fisher",
            desc=path2name)

# fetch significant pathways

test[test$hypergeometric<=0.05,]

# load the CORNA package

library(CORNA)

# read mirbase targets data for Mus musculus

targets <- miRBase2df.fun(file="arch.v5.txt.mus_musculus.zip")

# get transcripts associated with mmu-mir-708

tvec <- corna.map.fun(x=targets,
                      y="mmu-mir-708",
                      m="mir",
                      n="tran")

#Get links between transcript and GO term (molecular function) from BioMart:

tran2gomf  <- BioMart2df.fun(biomart="ensembl",  
                             dataset="mmusculus_gene_ensembl",
                             col.old=c("ensembl_transcript_id", "go_molecular_function_id"),
                             col.new=c("tran", "gomf"))

#Get links between GO id and term:

go2term <- GO2df.fun(url="ftp://ftp.ncbi.nlm.nih.gov/gene/DATA/gene2go.gz")

#population here is the set of all transcripts in tran2gomf

pvec <- as.vector(unique(tran2gomf[, "tran"]))

#sample here is the set of transcripts in both my.tran.vec and population:

samvec <- intersect(tvec, pvec) 

#Run CORNA test:

test <- corna.test.fun(samvec,
                       pvec,
                       tran2gomf,
                       fisher=T,
                       sort="fisher",
                       min.pop=5, 
                       desc=go2term) 

#significant list:

test[test$fisher <= 0.05, ]

# load the library and data

library(CORNA)
data(CORNA.DATA)

# the data used in this example comes from 
# http://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE7333
# here, three mice with mmu-mir-1-2 knocked out are compared
# to three wild type mice using the Affymetrix GeneChip Mouse Genome 
# 430 2.0 Array

# Zhao Y, Ransom JF, Li A, Vedantham V, von Drehle M, Muth AN, Tsuchihashi T, McManus MT, Schwartz RJ, 
# Srivastava D. Dysregulation of cardiogenesis, cardiac conduction, and cell cycle in mice lacking miRNA-1-2.
# Cell. 2007 Apr 20;129(2):303-17. 
# PMID: 17397913 

# this data can be read directly from GEO using a CORNA function:
# GEO.url    <- "ftp://ftp.ncbi.nih.gov/pub/geo/DATA/SOFT/by_series/GSE7333/GSE7333_family.soft.gz"
# microarray.df <- GEO2df.fun(url=GEO.url)

# however, we have packaged the data with CORNA so it is available using data(CORNA.DATA)
# we have also carried out a very simple linear model analysis to find 1273 differentially
# expressed affy probe ids.  This vector is called sam.probe.vec

# get a link from affy probe to ensembl transcript using BioMart2df.fun

tran2probe <- BioMart2df.fun(biomart="ensembl", 
                             dataset="mmusculus_gene_ensembl",
                             col.old=c("ensembl_transcript_id", "affy_mouse430_2"),
                             col.new=c("tran", "probe"))

# read mirbase targets data for Mus musculus

targets <- miRBase2df.fun(file="arch.v5.txt.mus_musculus.zip")

# get transcripts associated with mmu-mir-1

tvec <- corna.map.fun(x=targets,
                      y="mmu-mir-1",
                      m="mir",
                      n="tran")  

# link to significant transcripts  

sigprobe <- corna.map.fun(tran2probe,
                         tvec,
                         "tran",
                         "probe", all=TRUE)

# which ones are targeted and significant?

tsig <- intersect(sigprobe$probe, sam.probe.vec) 

# so the array data for these is:

my.microarray.df <- corna.sub.fun(x=microarray.df, 
                                  y=tsig)

#by default, draw lines for each of them:

corna.line.fun(x=my.microarray.df,
               array.order= c("GSM176756", "GSM176758", "GSM176759", "GSM176757", "GSM176760", "GSM176761"),
               las=0, ps=11, mar=c(7,3,3,1),
               main="Significant Probes")

#Draw barplots for the first 9 probes in a 3 times 3 figure:

corna.barplot.fun(x=my.microarray.df,
                  array.order= c("GSM176756", "GSM176758", "GSM176759", "GSM176757", "GSM176760", "GSM176761"),
                  array.colour=c(rep("black", 3), rep("white", 3)), 
                  row=3, 
                  column=3,
                  main="Significant probes targeted by mmu-mir-1",
                  line=3)

# draw all 45

corna.barplot.fun(x=my.microarray.df,
                  array.order= c("GSM176756", "GSM176758", "GSM176759", "GSM176757", "GSM176760", "GSM176761"),
                  array.colour=c(rep("black", 3), rep("white", 3)), 
                  row=7, 
                  column=7,
                  main="Significant probes targeted by mmu-mir-1",
                  line=3)