Absorb evidence into the model

Description

Absorb evidence into the model

Usage

AbsorbEvidence(tree, vars, values)

Arguments

Details

Absorb multiple types and pieces of evidences into a ClusterTree object. The discrete compartment of the ClusterTree will be automatically propagated after evidence absorption, so that the object will be ready for making queries and absorbing additional evidence.

Value

ClusterTree object with the evidence absorbed

Author(s)

Han Yu

References

Cowell, R. G. (2005). Local propagation in conditional Gaussian Bayesian networks. Journal of Machine Learning Research, 6(Sep), 1517-1550.

Lauritzen, S. L., & Spiegelhalter, D. J. (1988). Local computations with probabilities on graphical structures and their application to expert systems. Journal of the Royal Statistical Society. Series B (Methodological), 157-224.

Yu H, Moharil J, Blair RH (2020). BayesNetBP: An R Package for Probabilistic Reasoning in Bayesian Networks. Journal of Statistical Software, 94(3), 1-31. <doi:10.18637/jss.v094.i03>.

Examples

data(liver)
tree.init.p <- Initializer(dag=liver$dag, data=liver$data,
                           node.class=liver$node.class,
                           propagate = TRUE)
tree.post <- AbsorbEvidence(tree.init.p, c("Nr1i3", "chr1_42.65"), list(1,"1"))

An S4 class of the cluster tree.

Description

The ClusterTree object is the computational object for belief propagation.

Slots

cluster

A vector storing the name of clusters in the cluster tree.

node

A vector storing the name of nodes in the Bayesian network.

graph

A list of two graphNEL objects: $dag stores the graph of Bayesian network, $tree stores the graph of the cluster tree.

member

A named list of the node cluster membership.

parent

A named vector indicating the parent node of a given cluster in the cluster tree.

cluster.class

A named vector of logical values indicating whether a cluster is continuous or discrete.

node.class

A named vector of logical values indicating whether a node is continuous or discrete.

assignment

A named list indicating the assignment of discrete nodes discrete clusters.

propagated

A logical value indicating whether the discrete compartment has been propagated.

cpt

A named list of the conditional probability tables.

jpt

A named list of the joint distribution tables.

lppotential

A named list of the linear predictor potentials assigned to each cluster in the lppotential slots.

postbag

A named list of the linear predictor potentials assigned to each cluster in the postbag slots.

activeflag

A named vector of logical values indicating whether a continuous cluster is active.

absorbed.variables

A vector of characters indicating variables observed with hard evidence.

absorbed.values

A list indicating the values of the variables observed with hard evidence.

absorbed.soft.variables

A vector of characters indicating variables observed with soft or likelihood evidence.

absorbed.soft.values

A list of the likelihoods of the soft or likelihood evidence.

Compile the cluster tree

Description

Get the cluster sets and strong semi-elimination tree from the Bayesian network

Usage

ClusterTreeCompile(dag, node.class)

Arguments

Details

This function forms the cluster sets and the semi-elimination tree graph from the Bayesian network. The procedures include acquiring the elimination order, moralization, triangulation, obtaining cluster sets, forming strong elimination tree and strong semi-elimination tree. The cluster sets and the semi-elimination tree are required to initialize the cluster tree.

Value

tree.graph

a graphNEL object of semi-elimination tree.

dag

a graphNEL object of original Bayesian network.

cluster.sets

a list of members of each cluster.

node.class

a named vector of logical values, TRUE if node is discrete, FASLE if otherwise

elimination.order

a vector of node names sorted by the elimination order.

Author(s)

Han Yu

References

Cowell, R. G. (2005). Local propagation in conditional Gaussian Bayesian networks. Journal of Machine Learning Research, 6(Sep), 1517-1550.

Yu H, Moharil J, Blair RH (2020). BayesNetBP: An R Package for Probabilistic Reasoning in Bayesian Networks. Journal of Statistical Software, 94(3), 1-31. <doi:10.18637/jss.v094.i03>.

See Also

ElimTreeInitialize

Examples

data(liver)
cst <- ClusterTreeCompile(dag=liver$dag, node.class=liver$node.class)

Compute signed and symmetric Kullback-Leibler divergence

Description

Compute signed and symmetric Kullback-Leibler divergence of variables over a spectrum of evidence

Usage

ComputeKLDs(
  tree,
  var0,
  vars,
  seq,
  pbar = TRUE,
  method = "gaussian",
  epsilon = 10^-6
)

Arguments

Details

Compute signed and symmetric Kullback-Leibler divergence of variables over a spectrum of evidence. The signed and symmetric Kullback-Leibler divergence is also known as Jeffery's signed information (JSI) for continuous variables.

Value

a data.frame of the divergence

Author(s)

Han Yu

References

Cowell, R. G. (2005). Local propagation in conditional Gaussian Bayesian networks. Journal of Machine Learning Research, 6(Sep), 1517-1550.

Yu H, Moharil J, Blair RH (2020). BayesNetBP: An R Package for Probabilistic Reasoning in Bayesian Networks. Journal of Statistical Software, 94(3), 1-31. <doi:10.18637/jss.v094.i03>.

Examples

## Not run: 
data(liver)
tree.init.p <- Initializer(dag=liver$dag, data=liver$data,
                           node.class=liver$node.class,
                           propagate = TRUE)
klds <- ComputeKLDs(tree=tree.init.p, var0="Nr1i3",
                    vars=setdiff(tree.init.p@node, "Nr1i3"),
                    seq=seq(-3,3,0.5))
head(klds)

## End(Not run)

Initialize the elimination tree

Description

Initialize the elimination tree with the local models

Usage

ElimTreeInitialize(tree, dag, model, node.sets, node.class)

Arguments

Details

Initialize the elimination tree with the local models

Value

ClusterTree object with the local models incorporated

Author(s)

Han Yu

References

Cowell, R. G. (2005). Local propagation in conditional Gaussian Bayesian networks. Journal of Machine Learning Research, 6(Sep), 1517-1550.

Yu H, Moharil J, Blair RH (2020). BayesNetBP: An R Package for Probabilistic Reasoning in Bayesian Networks. Journal of Statistical Software, 94(3), 1-31. <doi:10.18637/jss.v094.i03>.

See Also

The functions ClusterTreeCompile and LocalModelCompile provide necessary objects to obtain ClusterTree object by initializing the elimination tree through this function.

Examples

data(liver)
cst <- ClusterTreeCompile(dag=liver$dag, node.class=liver$node.class)
models <- LocalModelCompile(data=liver$data, dag=liver$dag, node.class=liver$node.class)
tree.init <- ElimTreeInitialize(tree=cst$tree.graph,
                                dag=cst$dag,
                                model=models,
                                node.sets=cst$cluster.sets,
                                node.class=cst$node.class)

Queries of discrete variable distributions

Description

Obtain the joint, marginal, and conditional distributions of discrete variables

Usage

FactorQuery(tree, vars = c(), mode = c("joint", "conditional", "list"))

Arguments

Details

Query the joint distribution of any combination of discrete variables when mode is "joint", or conditional distribution of a discrete variable. The mode "list" return a list of variable combinations, such that joint distributions of any subset of them are ready for extraction. Queries outside this list are also supported but may take longer computing time. This function will also return marginal distribution if only one variable is queried.

Value

data.frame object specifying a joint or conditional distribution.

Author(s)

Han Yu

References

Cowell, R. G. (2005). Local propagation in conditional Gaussian Bayesian networks. Journal of Machine Learning Research, 6(Sep), 1517-1550.

Yu H, Moharil J, Blair RH (2020). BayesNetBP: An R Package for Probabilistic Reasoning in Bayesian Networks. Journal of Statistical Software, 94(3), 1-31. <doi:10.18637/jss.v094.i03>.

Examples

data(chest)
dag <- chest$dag
node.class <- rep(TRUE, length(dag@nodes))
names(node.class) <- dag@nodes
tree.init.p <- Initializer(dag=dag, data=chest$data,
                           node.class=node.class,
                           propagate=TRUE)
# joint distribution
FactorQuery(tree=tree.init.p, vars=c("tub", "xray", "dysp", "asia"), mode="joint")

# conditional distribution
FactorQuery(tree=tree.init.p, vars=c("xray"), mode="conditional")

Possible values of a discrete variable

Description

Obtain all the possible values of a discrete variable.

Usage

GetValue(tree, var, message = TRUE)

Arguments

Value

a vector of the possible values of discrete variable. If the variable is continuous, the returned value will be NULL.

Author(s)

Han Yu

Examples

data(toytree)
GetValue(toytree, "HDL")

Initialize a ClusterTree object

Description

Initialize a ClusterTree object

Usage

Initializer(dag, data, node.class, propagate = TRUE)

Arguments

Details

A wrapper function to initialize a ClusterTree object. It combines the functions of ClusterTreeCompile, LocalModelCompile, ElimTreeInitialize and Propagate, thus initialize the ClusterTree object in a single step.

Value

ClusterTree object

Author(s)

Han Yu

References

Cowell, R. G. (2005). Local propagation in conditional Gaussian Bayesian networks. Journal of Machine Learning Research, 6(Sep), 1517-1550.

Yu H, Moharil J, Blair RH (2020). BayesNetBP: An R Package for Probabilistic Reasoning in Bayesian Networks. Journal of Statistical Software, 94(3), 1-31. <doi:10.18637/jss.v094.i03>.

See Also

ClusterTreeCompile, LocalModelCompile, ElimTreeInitialize, Propagate

Examples

data(liver)
tree.init.p <- Initializer(dag=liver$dag, data=liver$data,
                           node.class=liver$node.class,
                           propagate = TRUE)

Model compilation

Description

Compile the local models

Usage

LocalModelCompile(data, dag = NULL, node.class = NULL)

Arguments

Details

This function compiles the local models, including the conditional probability tables for discrete variables, and linear predictor potentials for continuous variables. The qtlnet and qtl package need to be installed if data is a qtlnet object.

Value

pots

a list of discrete potentials (conditional probability tables) for each discrete variable.

bags

a list of sets of continuous potentials (lppotentials), each set for a continuous variables.

Author(s)

Han Yu

References

Cowell, R. G. (2005). Local propagation in conditional Gaussian Bayesian networks. Journal of Machine Learning Research, 6(Sep), 1517-1550.

Yu H, Moharil J, Blair RH (2020). BayesNetBP: An R Package for Probabilistic Reasoning in Bayesian Networks. Journal of Statistical Software, 94(3), 1-31. <doi:10.18637/jss.v094.i03>.

See Also

ElimTreeInitialize

Examples

data(liver)
models <- LocalModelCompile(data=liver$data, dag=liver$dag, node.class=liver$node.class)

Obtain marginal distributions

Description

Get the marginal distributions of multiple variables

Usage

Marginals(tree, vars)

Arguments

Details

Get the marginal distributions of multiple variables. The function Marginals returns a list of marginal distributions. The marginal distribution of a discrete variable is a named vector of probabilities. Meanwhile, the marginal distributions of continous variables in a CG-BN model are mixtures of Gaussian distributions. To fully represent this information, the marginal of a continuous variable is represented by a data.frame with three columns to specify parameters for each Gaussian distribution in the mixture, which are

mean

the mean value of a Gaussian distribution.

sd

the standard deviation of a Gaussian distribution.

n

the number of Gaussian mixtures

Value

marginals

a list of marginal distributions

types

a named vector indicating the types of the variables whose marginals are queried: TRUE for discrete, FALSE for continuous.

Author(s)

Han Yu

References

Cowell, R. G. (2005). Local propagation in conditional Gaussian Bayesian networks. Journal of Machine Learning Research, 6(Sep), 1517-1550.

Yu H, Moharil J, Blair RH (2020). BayesNetBP: An R Package for Probabilistic Reasoning in Bayesian Networks. Journal of Statistical Software, 94(3), 1-31. <doi:10.18637/jss.v094.i03>.

See Also

PlotMarginals for visualization of the marginal distributions, SummaryMarginals for summarization of the marginal distributions of continuous variables.

Examples

data(liver)
tree.init.p <- Initializer(dag=liver$dag, data=liver$data,
                           node.class=liver$node.class,
                           propagate = TRUE)
tree.post <- AbsorbEvidence(tree.init.p, c("Nr1i3", "chr1_42.65"), list(1,"1"))
marg <- Marginals(tree.post, c("HDL", "Ppap2a"))
marg$marginals$HDL
head(marg$marginals$Ppap2a)

Plot the Bayesian network

Description

Plot and compare two Bayesian networks with different evidence(s) absorbed and propagated.

Usage

PlotCGBN(
  tree.1,
  tree.2,
  fontsize = NULL,
  pbar = FALSE,
  plotting = TRUE,
  epsilon = 10^-6
)

Arguments

Details

Network visualization of the node-specific differences between Bayesian Networks with the same topology, but evidence that has been absorbed and propagated. The change of marginal distribution of each node is measured by signed and symmetric Kullback-Leibler divergence. The sign indicates the direction of change, with tree.1 considered as the baseline. The magnitude of the change is reflected by the value. Nodes that are white are d-separated from the evidence. This function requires Rgraphviz package.

Value

a plot of Bayesian network

a vector of signed symmetric Kullback-Leibler divergence

Author(s)

Han Yu

References

Cowell, R. G. (2005). Local propagation in conditional Gaussian Bayesian networks. Journal of Machine Learning Research, 6(Sep), 1517-1550.

Yu H, Moharil J, Blair RH (2020). BayesNetBP: An R Package for Probabilistic Reasoning in Bayesian Networks. Journal of Statistical Software, 94(3), 1-31. <doi:10.18637/jss.v094.i03>.

Examples

## Not run: 
library("Rgraphviz")
data(toytree)
tree.post <- AbsorbEvidence(toytree, c("Nr1i3"), list(1))
PlotCGBN(tree.1=toytree, tree.2=tree.post)

## End(Not run)

Plot the marginal distributions

Description

Plot the marginal distributions.

Usage

PlotMarginals(marginals, groups = NULL)

Arguments

Details

Plot the marginal distributions. Marginals of discrete variables are plotted as bar plots, while those of continuous variables as density plots.

Author(s)

Han Yu

References

Cowell, R. G. (2005). Local propagation in conditional Gaussian Bayesian networks. Journal of Machine Learning Research, 6(Sep), 1517-1550.

See Also

Marginals

Examples

data(toytree)
marg <- Marginals(toytree, c("Neu1", "Nr1i3", "chr1_42.65", "Spgl1"))
PlotMarginals(marginals=marg, groups=NULL)

Plot the cluster tree

Description

Plot the structure of a ClusterTree object

Usage

PlotTree(tree, color = "gray90")

Arguments

Details

Plot the structure of clustertree object, with the nodes labeled by corresponding elimination node. The circles represent continuous clusters, while the boxes represent discrete clusters. This function requires Rgraphviz package.

Author(s)

Han Yu

References

Cowell, R. G. (2005). Local propagation in conditional Gaussian Bayesian networks. Journal of Machine Learning Research, 6(Sep), 1517-1550.

Yu H, Moharil J, Blair RH (2020). BayesNetBP: An R Package for Probabilistic Reasoning in Bayesian Networks. Journal of Statistical Software, 94(3), 1-31. <doi:10.18637/jss.v094.i03>.

Cowell, R. G. (2005). Local propagation in conditional Gaussian Bayesian networks. Journal of Machine Learning Research, 6(Sep), 1517-1550.

Examples

## Not run: 
library("Rgraphviz")
data(toytree)
PlotTree(toytree)

## End(Not run)

Propagate the cluster tree

Description

This function propagates the discrete compartment of a ClusterTree object.

Usage

Propagate(tree, targets = NA)

Arguments

Details

The discrete compartment must be propagted to get the joint distributions of discrete variables in each discrete clusters. A ClusterTree object must be propagated before absorbing evidence and making queries.

Value

a ClusterTree object

References

Cowell, R. G. (2005). Local propagation in conditional Gaussian Bayesian networks. Journal of Machine Learning Research, 6(Sep), 1517-1550.

Lauritzen, S. L., & Spiegelhalter, D. J. (1988). Local computations with probabilities on graphical structures and their application to expert systems. Journal of the Royal Statistical Society. Series B (Methodological), 157-224.

Yu H, Moharil J, Blair RH (2020). BayesNetBP: An R Package for Probabilistic Reasoning in Bayesian Networks. Journal of Statistical Software, 94(3), 1-31. <doi:10.18637/jss.v094.i03>.

Examples

data(liver)
tree.init <- Initializer(dag=liver$dag, data=liver$data,
                         node.class=liver$node.class,
                         propagate = FALSE)
tree.init@propagated
tree.init.p <- Propagate(tree.init)
tree.init.p@propagated

Sampling from the Bayesian network

Description

Sampling from the joint distribution of all applicable nodes in the Bayesian network.

Usage

Sampler(tree, n)

Arguments

Value

a dataframe of generated data

Author(s)

Han Yu

References

Cowell, R. G. (2005). Local propagation in conditional Gaussian Bayesian networks. Journal of Machine Learning Research, 6(Sep), 1517-1550.

Yu H, Moharil J, Blair RH (2020). BayesNetBP: An R Package for Probabilistic Reasoning in Bayesian Networks. Journal of Statistical Software, 94(3), 1-31. <doi:10.18637/jss.v094.i03>.

Examples

data(toytree)
Sampler(tree = toytree, n = 10)

Summary a continuous marginal distribution

Description

This function summary the marginal distributions of continuous variables by outputing the mean, standard deviation, and number of subpopulations

Usage

SummaryMarginals(marginals)

Arguments

Value

a data.frame object containing information about the marginal distributions for continuous variables. The marginal distributions of continous variables in a CG-BN model are mixtures of Gaussian distributions. Therefore, besides the mean and standard deviation, the object has an additional column to specify the number of Gaussian mixtures.

mean

the mean value of a Gaussian distribution.

sd

the standard deviation of a Gaussian distribution.

n

the number of Gaussian distributions in the mixture.

References

Cowell, R. G. (2005). Local propagation in conditional Gaussian Bayesian networks. Journal of Machine Learning Research, 6(Sep), 1517-1550.

Yu H, Moharil J, Blair RH (2020). BayesNetBP: An R Package for Probabilistic Reasoning in Bayesian Networks. Journal of Statistical Software, 94(3), 1-31. <doi:10.18637/jss.v094.i03>.

See Also

Marginals

Examples

data(liver)
tree.init.p <- Initializer(dag=liver$dag, data=liver$data,
                           node.class=liver$node.class,
                           propagate = TRUE)
marg <- Marginals(tree.init.p, c("HDL", "Ppap2a", "Neu1"))
SummaryMarginals(marginals=marg)

Convert a bn object to graphNEL object

Description

Convert a bn object to graphNEL object while removing isolated nodes

Usage

bn_to_graphNEL(graph_bn)

Arguments

Value

a graphNEL object

Author(s)

Han Yu

A simulated data from the Chest Clinic example

Description

Simulated data from the Chest Clinic example (also known as the Asia example) from Lauritzen and Spiegelhalter, 1988.

Usage

data(chest)

Format

The data set chest contains two objects:

data

a data.frame object of 10000 observations and 8 discrete variables.

dag

a graphNEL object specifying the network structure.

References

Lauritzen and Spiegelhalter (1988) Local Computations with Probabilities on Graphical Structures and their Application to Expert Systems (with Discussion). J. Roy. Stat. Soc. 50, p. 157-224.

Dethlefsen, C., & Hojsgaard, S. (2005). A common platform for graphical models in R: The gRbase package. Journal of Statistical Software, 14(17), 1-12.

A ClusterTree Example of Emission Model

Description

A propagated ClusterTree object named emission. This model contains nine variables, indlucing three discrete: Filter State (Fs), Waste Type (W), Burning Regimen (B) and six continuous variables: Metals in Waste (Min), Metals Emission (Mout), Filter Efficiency (E), Dust Emission (D), CO2 Concentration in Emission (C), Light Penetrability (L).

Usage

data(emission)

Format

The data set contains a propagated ClusterTree object emission ready for evidence absorption and making queries.

References

Lauritzen, Steffen L., and Frank Jensen. Stable local computation with conditional Gaussian distributions. Statistics and Computing 11.2 (2001): 191-203.

A simulated data from the Emission example

Description

Simulated data from the Emission example (also known as the Waste Incinerator example)

Usage

data(emission1000)

Format

The data set emission1000 contains two objects:

data

a data.frame object of 1000 observations and 3 discrete variables and 6 continuous variables.

dag

a graphNEL object specifying the network structure.

References

Lauritzen, S. L., & Jensen, F. (2001). Stable local computation with conditional Gaussian distributions. Statistics and Computing, 11(2), 191-203.

Mus Musculus HDL QTL data from Leduc et. al. (2012)

Description

Liver QTL data was obtained from a F2 inner-cross between inbred MRL/MpJ and SM/J strains of mice.

Usage

data(liver)

Format

The data set liver contains three objects: the data, a learned Bayesian network structure and vector specifying node type. The fields are described as follows:

data

a data.frame object that contains 280 samples (rows) and 15 variables: genotype data (genotype states at 5 SNP markers) and phenotype data (HDL levels and normalized expression values of 10 genes). Three of these phenotypes are dichotomized, including Cyp2b10, Spgl1 and HDL. Genotypes and dichotomized phenotypes are of class factor and continuous phenotypes are of class numeric.

dag

a graphNEL object, which is the network structure learned by qtlnet package.

node.class

a named vector of logical values indicating whether each node is discrete.

References

Leduc MS, Blair RH, Verdugo RA, Tsaih SW, Walsh K, Churchill GA, Paigen B.(2012). "Using bioinformatics and systems genetics to dissect HDL-cholesterol genetics in an MRL/MpJ x SM/J intercross." J Lipid Res., 6, 1163-75.

Convert qtlnet to graphNEL object

Description

Extract network structure from qtlnet object and convert to graphNEL object

Usage

qtlnet_to_graphNEL(data)

Arguments

Details

This function extracts network structure from qtlnet object and convert to graphNEL object. The example data can be downloaded from <https://github.com/hyu-ub/BayesNetBP>.

Value

Author(s)

Han Yu

Examples

## Not run: 
load(liverqtl.rda)
qtlnet_to_graphNEL(liverqtl$qtlnet.fit)

## End(Not run)

Launch the BayesNetBP Shiny App

Description

Launch the BayesNetBP Shiny App

Usage

runBayesNetApp(launch.browser = TRUE)

Arguments

Details

The function runBayesNetApp lauches the Shiny App accompanied with this package. The app loads the toytree example by default and allows users to load customized ClusterTree object. In order to use this feature, a ClusterTree object should be built, propagated and named tree.init.p, and then saved as a .RDATA file. This file can be read in by the app.

The console of BayesNetBP Shiny App comprises three panels. The first part controls the model loading, visualization and subnetwork selection. The Fit function fits the entire graph in the window. The Fit Selected function fits the selected subnetwork to the window. The user can subset the network for visualization. The Expand function can trace the one hop neighbor of selected nodes in a stepwise manner. After selecting desired node sets, the user can subset the graph by the Subset function.

The second panel is used for absorption of fixed and hard evidences. The users can add multiple pieces of evidence to a list and absorb them into the model simultaneously. Marginals of other nodes can be quried as density or bar plots by node types. If a set of evidence has been absorbed, the marginals both before and after absorption will be returned to facilitate comparison. To query the marginals, the user can select the node of interest in the graph, and then click Marginal of Selected. The Shift in Marginals function computes the signed and symmetric Kullback-Liebler divergence for all applicable nodes in the network, and colors the nodes by their divergence and change in directions.

The function for systematic assessment of variable marginal shifts is provided in the third panel. It allows user to specify which node to absorb the spectrum of evidence in the select menu and click Select Observed, and to select whose divergence to be calculated by selecting the node in the menu and then clicking Add to Plot. Alternatively, the user can use Add All function to select all applicable nodes into the plotting list. The result is visualized in an interactive plot. The Min, Max and Step controls the range of values of the evidence to be absorbed.

Author(s)

Han Yu

References

Yu H, Moharil J, Blair RH (2020). BayesNetBP: An R Package for Probabilistic Reasoning in Bayesian Networks. Journal of Statistical Software, 94(3), 1-31. <doi:10.18637/jss.v094.i03>.

Yu H, Moharil J, Blair RH (2020). BayesNetBP: An R Package for Probabilistic Reasoning in Bayesian Networks. Journal of Statistical Software, 94(3), 1-31. <doi:10.18637/jss.v094.i03>.

Examples

## Not run: 
# load or install required packages to run App
library("shiny")
library("googleVis")
library("devtools")
devtools::install_github("cytoscape/cyjShiny")
library("cyjShiny")
# run the App in browser
runBayesNetApp(launch.browser=TRUE)

## End(Not run)

A ClusterTree Example of Liver Model

Description

A propagated ClusterTree object named toytree, obtained from liver QTL data.

Usage

data(toytree)

Format

The data set contains a propagated ClusterTree object toytree, which is ready for evidence absorption and making queries.

Saccharomyces Cerevisiae eQTL data from Kruglak et. al. (2005)

Description

eQTL data from 112 F1 segregants from a cross between BY4716 and RM11-1a strains of Saccharomyces Cerevisiae.

Usage

data(yeast)

Format

The data set yeast is a data frame of 112 observations of 50 variables: genotype data (genotype states at 12 SNP markers) and phenotype data (normalized and discretized expression values of 38 genes). Both genotypes and phenotypes are of class factor.

Details

The yeast dataset is a subset of the widely studied yeast expression dataset comprising of 112 F1 segregants from a cross between BY4716 and RM11-1a strains of Saccharomyces Cerevisiae. The original dataset consists of expression values reported as log2(sample/ BY reference) for 6216 genes. The data can be accessed in Gene Expression Omnibus (GEO) by accession number (GSE1990). After linkage analysis and filtering based on location and significance of QTL, a final set of 38 genes and their corresponding 12 SNP markers were identified and included in the yeast dataset. The gene expression values are discretized around the median and have two states, 1 (above or equal to median) and -1 (below median). re are two genotype states: 1 or 2. Thus the final dataset is a data frame of 112 observations (genotype) of 12 variables (SNP markers) and normalized gene expression of 38 variables (genes).

References

Brem RB, Kruglyak L. The landscape of genetic complexity across 5,700 gene expression traits in yeast. Proc Natl Acad Sci U S A 2005 Feb 1;102(5):1572-7.

Brem RB, Storey JD, Whittle J, Kruglyak L. Genetic interactions between polymorphisms that affect gene expression in yeast. Nature 2005 Aug 4;436(7051):701-3.