ARACNe-AP / Wiki / Home

Juan Manuel Luque - 2018-05-08

Dear Federico,Alexander and Aracne users in general:

Please, let me I write you about Aracne-geWorkbench software because I don't know whom I must ask or write to solve my problems with this software. I am a new Aracne-geWorkbench user and I am trying to do an Aracne analysis from geWorkbench platform with a list of 368 genes I interest to study and visualize graphically the possible networks between these genes.But once I load my list of genes and my list of hub genes I see the next message/error on the screen:
"Exception happened in aracne computation: org.geworkbench.components.aracne.AracneException: There is no microarray name provided".

And I don t know where (what field) I must to fill or correct the microarray name, and so finally to get the Aracne analysis for my list of genes.
Please if you could help me or tell me whom I should write or ask to get help and solve this problem ,I'll thank you very much.
Best regards
Juan Manuel Luque

If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
- Federico Giorgi - 2018-05-20
  
  Hi Juan!
  Aracne-geWrokbench is not maintained by us, you should contact Aristidis Floratos at Columbia University for it. As an alternative, you can use our Aracne-AP, which doesn't require a microarray name to run!
  
  If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
  - Juan Manuel Luque - 2018-05-20
    
    Thank you very much
    
    Juan Manuel
    
    ----Original Message----
    From: fgiorgi@users.sourceforge.net
    Date: 20/05/2018 12:26
    To: "[aracne-ap:wiki] "Home@wiki.aracne-ap.p.re.sourceforge.net
    Subj: [aracne-ap:wiki] Re: Discussion for Home page
    
    Hi Juan!
    Aracne-geWrokbench is not maintained by us, you should contact
    Aristidis Floratos at Columbia University for it. As an alternative,
    you can use our Aracne-AP, which doesn't require a microarray name to
    run!
    
    Sent from sourceforge.net because you indicated interest in https://sourceforge.net/p/aracne-ap/wiki/Home/
    
    To unsubscribe from further messages, please visit https://sourceforge.net/auth/subscriptions/
    
    If you would like to refer to this comment somewhere else in this project, copy and paste the following link:

Juan Manuel Luque - 2018-05-21

Hi Federico,
I am interested to use AracneAP.I have downloaded it, but I can´t open the file downloaded.Please, could you help me to install AracneAP in my computer?.Could you send me the whole steps to install AracneAP in my computer (Windows 7 and also Ubuntu-16 in other computer)?If I have installed Aracne-geWorkbench do I need to install AracneAP too?
Waiting for your answer, please, thank you very much for your help and best regards.
Juan Manuel Luque

If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
- Federico Giorgi - 2018-09-27
  
  Hi Juan!
  This implementation of aracne is to be run directly as a jar file from the command line. E.g. java -jar Aracne.jar
  Try to run some of the examples in the wiki from the command line (from the Windows command prompt or the Ubuntu terminal, for example).
  
  If you would like to refer to this comment somewhere else in this project, copy and paste the following link:

Xiaoyong Fu - 2020-06-19

Hello, I am trying to run: java -Xmx5G -jar Aracne.jar ... in my mac terminal, but got the error:
Exception in thread "main" java.lang.NoClassDefFoundError: jargs/gnu/CmdLineParser$OptionException
at java.base/java.lang.Class.forName0(Native Method)
at java.base/java.lang.Class.forName(Class.java:427)
at org.eclipse.jdt.internal.jarinjarloader.JarRsrcLoader.main(JarRsrcLoader.java:56)
Caused by: java.lang.ClassNotFoundException: jargs.gnu.CmdLineParser$OptionException
at java.base/java.net.URLClassLoader.findClass(URLClassLoader.java:435)
at java.base/java.lang.ClassLoader.loadClass(ClassLoader.java:589)
at java.base/java.lang.ClassLoader.loadClass(ClassLoader.java:522)
... 3 more
I have the expression file and gene list and the Aracne.jar under the same folder. Please help me to address this issue.
Thanks a lot!
Xiaoyong

If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
- Claire Rioualen - 2020-08-07
  
  Hi, I'm hitting the same issue, were you able to solve it?
  
  If you would like to refer to this comment somewhere else in this project, copy and paste the following link:

Xiaoyong Fu - 2020-08-07

I think I ended up using a HPC cluster to run Java and it works. It may need more computering power which can not be afforded by my personal Mac. thanks.

If you would like to refer to this comment somewhere else in this project, copy and paste the following link:

Hi, thank you for your work building ARACNe-AP. I installed the software on my local machine using the command line and was able to run the program using the matrix.txt and tfs.txt files as an example and generated the network.txt file. I would like to ultimately use VIPER in R to predict protein activity but the input to VIPER is the .adj output file from ARACNe along with the expression matrix used to generate the .adj file. I am having trouble figuring out how to generate the .adj file. How would I go about doing this? Thank you for your time.

Hi there! Thanks for using ARACNe-AP. To convert the ARACNe-AP into an R object, this is what we used to do. Assuming all the outputs of ARACNe-AP are saved as files 3col.gz

Loading code:

library(viper)
library(mixtools)

Necessary functions:

aracne2regulon<-function(afile, eset) {
    dset <- eset
    annot <- rownames(eset)
    names(annot) <- rownames(eset)
    rm(eset)

    tmp <- as.matrix(read.delim(afile,as.is=TRUE,header=TRUE)[,1:3])
    aracne <- data.frame(tf=tmp[, 1], target=tmp[, 2], mi=as.numeric(tmp[, 3])/max(as.numeric(tmp[, 3])))

    tmp <- aracne[!is.na(aracne$mi), ]
    tmp <- tmp[rowSums(matrix(as.matrix(tmp[, 1:2]) %in% rownames(dset), nrow(tmp), 2))==2, ]
    aracne <- tapply(1:nrow(tmp), tmp$tf, function(pos, tmp) {
        tfmode <- rep(0, length(pos))
        names(tfmode) <- tmp$target[pos]
        list(tfmode=tfmode, likelihood=tmp$mi[pos])
    }, tmp=tmp)
    names(aracne) <- levels(tmp$tf)
    aracne <- TFmode1(aracne, dset)
    rm(dset)

    # removing missing data from the aracne regulon
    aracne <- aracne[names(aracne) != "NA"]
    aracne <- lapply(aracne, function(x) {
        filtro <- !(names(x$tfmode)=="NA" | is.na(x$tfmode) | is.na(x$likelihood))
        x$tfmode <- x$tfmode[filtro]
        x$likelihood <- x$likelihood[filtro]
        return(x)
    })
    aracne <- aracne[sapply(aracne, function(x) length(names(x$tfmode)))>0]
    regul <- TFscore(aracne)
    class(regul) <- "regulon"
    return(regul)
}

TFmode1 <- function (regulon, expset, method = "spearman") {
    expset <- filterCV(expset)
    regulon <- updateRegulon(regulon)
    regulon <- regulon[names(regulon) %in% rownames(expset)]
    regulon <- lapply(regulon, function(x, genes) {
        filtro <- names(x$tfmode) %in% genes
        x$tfmode <- x$tfmode[filtro]
        if (length(x$likelihood) == length(filtro)) 
            x$likelihood <- x$likelihood[filtro]
        return(x)
    }, genes = rownames(expset))
    tf <- unique(names(regulon))
    tg <- unique(unlist(lapply(regulon, function(x) names(x$tfmode)), use.names = FALSE))
    cmat <- cor(t(expset[rownames(expset) %in% tf, ]), t(expset[rownames(expset) %in% tg, ]), method = method)
    reg <- lapply(1:length(regulon), function(i, regulon, cmat) {
        tfscore <- cmat[which(rownames(cmat) == names(regulon)[i]), match(names(regulon[[i]]$tfmode), colnames(cmat))]
        list(tfmode = tfscore, likelihood = regulon[[i]]$likelihood)
    }, regulon = regulon, cmat = cmat)
    names(reg) <- names(regulon)
    return(reg)
}

TFmode2 <- function (regulon, expset, regexpset, method = "spearman") {
    tf <- unique(names(regulon))
    tg <- unique(unlist(lapply(regulon, function(x) names(x$tfmode)), use.names = FALSE))
    cmat <- cor(t(regexpset[rownames(regexpset) %in% tf, ]), t(expset[rownames(expset) %in% tg, ]), method = method)
    reg <- lapply(1:length(regulon), function(i, regulon, cmat) {
        tfscore <- cmat[which(rownames(cmat) == names(regulon)[i]), match(names(regulon[[i]]$tfmode), colnames(cmat))]
        list(tfmode = tfscore, likelihood = regulon[[i]]$likelihood)
    }, regulon = regulon, cmat = cmat)
    names(reg) <- names(regulon)
    return(reg)
}

TFscore <- function (regul, mu = NULL, sigma = NULL, verbose=TRUE) {
    if (length(mu) == 3 & length(sigma) == 3)
        fit <- list(mu = mu, sigma = sigma)
    else {
        tmp <- unlist(lapply(regul, function(x) x$tfmode), use.names = FALSE)
        fit <- list(mu = c(-0.5, 0, 0.5), sigma = c(0.15, 0.25, 0.15), lambda = c(0.2, 0.4, 0.4), all.loglik = rep(0, 1001))
        count <- 0
        while (length(fit$all.loglik) > 1000 & count < 3) {
            fit <- normalmixEM(tmp, mu = fit$mu, sigma = fit$sigma, lambda = fit$lambda, mean.constr = c(NA, 0, NA), maxit = 1000, verb = FALSE)
            count <- count + 1
        }
    }
    if (verbose) message("mu: ", paste(fit$mu, collapse = ", "), ". sigma: ", paste(fit$sigma, collapse = ", "))
    regul <- lapply(regul, function(x, fit) {
        g2 <- pnorm(x$tfmode, fit$mu[3], fit$sigma[3], lower.tail = TRUE)
        g1 <- pnorm(x$tfmode, fit$mu[1], fit$sigma[1], lower.tail = FALSE)
        g0 <- pnorm(x$tfmode, fit$mu[2], fit$sigma[2], lower.tail = FALSE)
        g00 <- pnorm(x$tfmode, fit$mu[2], fit$sigma[2], lower.tail = TRUE)
        x$tfmode <- g2/(g1 + g0 + g2) * (x$tfmode >= 0) - g1/(g1 + g00 + g2) * (x$tfmode < 0)
        return(x)
    }, fit = fit)
    return(regul)
}

updateRegulon <- function(regul) {
    if (is.null(names(regul[[1]]))) {
        tmp <- lapply(regul, function(x) {
            tmp <- rep(0, length(x))
            names(tmp) <- x
            list(tfmode=tmp, likelihood=rep(1, length(tmp)))
        })
        return(tmp)
    }
    if (names(regul[[1]])[1]=="tfmode") return(regul)
    return(lapply(regul, function(x) list(tfmode=x, likelihood=rep(1, length(x)))))
}

Actual execution. You need 1) the network as network.3col.gz and 2) the expression matrix in R format expmat.rda from which you generated the network

load("expmat.rda")
regul<-aracne2regulon(afile="network.3col.gz",eset=expmat)
class(regul)<-"regulon"
save(regul,file="network.rda")

Hope this helps. Let me know if it works!

ARACNe-AP Wiki

Network Reverse Engineering through AP inference of Mutual Information

Home

Overview

Input files needed to run ARACNe

Steps required to run ARACNe

Optional ways to run ARACNe

Output of ARACNe-AP

Input file format

Gene lists

Dataset

Parameters

Examples

Example 1 – calculate threshold with a fixed seed

Example 2 – run ARACNe on a single bootstrap

Example 3 – run 100 reproducible bootstraps

Example 4 – consolidate bootstraps in the output folder

Example 5 – run a single ARACNE with no bootstrap and no DPI

Example 6 – consolidate bootstraps without Bonferroni correction

License

Project Members:

Discussion