bobthecat

library(GEOquery); library(RankProd); library(mouse4302.db)
## Download the data from GEO
gse12499 <- getGEO('GSE12499',GSEMatrix=TRUE)
e <- exprs(gse12499[[1]])
dim(e)
[1] 45101    10

bobthecat

library(twitteR)
library(ggplot2)
library(grid)

# download all that you can
pol <- userTimeline('BeijingAir', n=3200)
length(pol)
# 3200

myGrep <- function(x){

bobthecat

library(twitteR)
library(ggplot2)
library(grid)

# download all that you can
pol <- userTimeline('BeijingAir', n=3200)
length(pol)
# 3200

myGrep <- function(x){

bobthecat

library(rbenchmark)
res <- benchmark( cosine(mat),
                  cosineRcpp(mat),
                  columns=c("test", "replications", "elapsed",
                           "relative", "user.self", "sys.self"),
                  order="relative",
                  replications=1000)

res
             test replications elapsed relative user.self sys.self

bobthecat

require(inline)
require(RcppArmadillo)

## extract cosine similarity between columns
cosine <- function(x) {
    y <- t(x) %*% x
    res <- 1 - y / (sqrt(diag(y)) %*% t(sqrt(diag(y))))
    return(res)
}

bobthecat

library("igraph")
gg <- graph.data.frame(ppi)

plot(gg,
  layout = layout.fruchterman.reingold,
  vertex.label = V(gg)$name,
  vertex.label.color= "black",
  edge.arrow.size=0,
  edge.curved=FALSE
)

bobthecat

ppi <- get.ppiNCBI(head(glist, 20))
[1] "7 interactions found"
## Annotate the gene list with Mus musculus metadata
library(org.Mm.eg.db)
ppi$egSymbol <- mget(ppi$egID, envir=org.Mm.egSYMBOL, ifnotfound=NA)
ppi$intID <- mget(ppi$intSymbol, envir=org.Mm.egSYMBOL2EG, ifnotfound=NA)
ppi <- ppi[,c(3,2,1,4)]
ppi
  egSymbol intSymbol  egID  intID
1  Ifi202b    Pou5f1 26388  18999

bobthecat

get.ppiNCBI <- function(g.n) {
  require(XML)
  ppi <- data.frame()
  for(i in 1:length(g.n)){
    o <- htmlParse(paste("http://www.ncbi.nlm.nih.gov/gene/", g.n[i], sep=''))
    # check if interaction table exists
    exist <- length(getNodeSet(o, "//table//th[@id='inter-prod']"))>0
    if(exist){
      p <- getNodeSet(o, "//table")
      ## need to know which table is the good one

bobthecat

library(org.Mm.eg.db); library(gplots); library(bioDist)

# get gene in GO cat
gGOcat <- list()
for(i in 1:dim(mrnaGO)[1]){
  gGOcat[[mrnaGO$GOBPID[i]]] <- as.vector(unlist(mget(mrnaGO$GOBPID[i], org.Mm.egGO2ALLEGS)))
}

# filter each GO cat by the genes found to be DE
f <- function(vecGO, vecDE){

bobthecat

source('source_https.r')
## You can now source R scripts from GitHub. The RAW URL is needed.
source_https('https://raw.github.com/bobthecat/codebox/master/GO_over.r')

## Define the universe
library(mouse4302.db)
uniqueId <- unique(as.vector(unlist(as.list(mouse4302ENTREZID))))
entrezUniverse <- uniqueId[!is.na(uniqueId)]
length(entrezUniverse)
[1] 20877