Coexpression EC2 Set-Up Details

These notes describe how we configured a custom AMI for running Coexpression on EC2.

The template ec2 instance (from which AMI is to be generated): ec2-184-73-105-132.compute-1.amazonaws.com

Note:  When ssh-ing to an instance, connect as 'root' user.

Set-up steps:

Created EC2 instance using BioConductor AMI. Then:

install.packages(c("MASS", "class", "MVA", "cluster", "survival", "rpart", "lattice", "scatterplot3d", "impute", "Hmisc"))

cran mirror 81

install.packages(c("RCurl", "bitops", "rjson", "RUnit"))

also need "sma" which is available here:
http://cran.r-project.org/src/contrib/Archive/sma/
... but it won't build (!)
... so we got the version from Belltown and put it here
"/usr/local/lib64/R/library"

Note: We also installed the WGCNA package from UCLA and a 'packaged' version of Bin Zhang's coexpression code.

copied
C:\Users\bhoff\eclipse projects\trunk\client\rSynapseClient
to
/usr/local/ec2-home

in /usr/local/ec2-home
R CMD INSTALL rSynapseClient

Started R and ran this script:

require(synapseClient)
require()
synapseAuthServiceEndpoint("https://staging-auth.elasticbeanstalk.com/auth/v1")
synapseRepoServiceEndpoint("https://staging-reposervice.elasticbeanstalk.com/repo/v1")
synapseLogin('matt.furia@sagebase.org','furia' (mailto:%27matt.furia@sagebase.org%27,%27furia%27))

datasets <- getDatasets()
dsId <- datasets$dataset.id[PLFM:datasets$dataset.name == "Harvard Brain Tissue Resource Center"]
ds <- getDataset(dsId)
layers <- getDatasetLayers(ds)

## hack
## get the expression layer and load it
layers <- layers$results
layerType <- NULL
for(i in 1:length(layers)){
layerTypei <- layers[PLFM:i]$type
}

layers <- layers[PLFM:which(layerType == "E")]
for(i in 1:length(layers)){
#look for gene expression layer
indx <- grep("Agilent 44K",layers[PLFM:i]$platform)
if(length(indx) == 0) next
break
}
layers <- layers[[PLFM:i]]
exprLayerFiles <- synapseClient:::.cacheFiles(entity=layers)

synapseClient:::.intergrationTest()