Data: Sample Data

In ptycho: Bayesian Variable Selection with Hierarchical Priors

Description

Data objects used in examples in this package.

Usage

data(tinysim)
data(ptychoIn)
data(ptychoOut)

Format

The object tinysim is an object returned by createDataBayesModel with mode equal to “pleiotropy”.

The object ptychoIn is an object returned by createDataBayesModel with mode equal to “gene”.

The object ptychoOut is an object returned by ptycho applied to the data in ptychoIn using the Across Sites prior.

Details

These data objects are constructed to illustrate certain features in the examples while still being small enough not to be burdensome.

The object tinysim contains simulated data. Its design matrix is 100-by-10. It has 10 replicates, each with a 100-by-5 response matrix. It is generated by createPubData.

The object ptychoIn also contains simulated data generated by createPubData. Its design matrix is 3000-by-10 because, for its effect sizes, n must be about that large to distinguish signal from noise as explained in the supplemental text to Stell and Sabatti (2015). To keep the object small, it has only 1 replicate, which has only one response.

The object ptychoOut is generated by

    G <- 2; p <- ncol(ptychoIn$X)
    groups <- createGroupsSim(G, p)
    state <- list(list(indic.grp=rep(FALSE,G),
                       indic.var=matrix(FALSE,nrow=p,ncol=1), tau=1),
                  list(indic.grp=rep(TRUE,G),
                       indic.var=matrix(TRUE,nrow=p,ncol=1), tau=1))
    ptychoOut <- ptycho(X=ptychoIn$X, y=ptychoIn$replicates[[1]]$y,
                        groups=groups, initStates=state,
                        only.means=10000*seq_len(5), random.seed=12345)
  

Source

Stell, L. and Sabatti, C. (2015) Genetic variant selection: learning across traits and sites, arXiv:1504.00946.

See Also

createPubData, ptycho

ptycho documentation built on May 2, 2019, 9:45 a.m.