Package 'pvclust' - R

Copy and paste this link to your website, so they can see this document directly without any plugins.



Keywords

lung, data, been, data(lung), original, Description, genes, cluster, Ryota, each, Resampling, Bootstrap, Multiscale, Suzuki, Microarray, 2015-10-23, License, analysis., topics, provides, 16:23:16, Date/Publication, (approximately, CRAN, unbiased), Repository, p-value, well, (bootstrap, probability)

Transcript

Package ‘pvclust’
October 23, 2015
Version 2.0-0
Date 2015-10-23
Title Hierarchical Clustering with P-Values via Multiscale Bootstrap
Resampling
Author Ryota Suzuki , Hidetoshi Shimodaira

Maintainer Ryota Suzuki
Depends R (>= 2.10.0)
Suggests MASS, parallel
Description An implementation of multiscale bootstrap resampling for
assessing the uncertainty in hierarchical cluster analysis.
It provides AU (approximately unbiased) p-value as well as
BP (bootstrap probability) value for each cluster in a dendrogram.
License GPL (>= 2)
URL http://www.sigmath.es.osaka-u.ac.jp/shimo-lab/prog/pvclust/
NeedsCompilation no
Repository CRAN
Date/Publication 2015-10-23 16:23:16
R topics documented:
lung . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
msfit . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
msplot . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
plot.pvclust . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
print.pvclust . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
pvclust . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
pvpick . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
seplot . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
Index 14
1
2 lung
lung DNA Microarray Data of Lung Tumors
Description
DNA Microarray data of 73 lung tissues including 67 lung tumors. There are 916 observations of
genes for each lung tissue.
Usage
data(lung)
Format
data frame of size 916× 73.
Details
This dataset has been modified from original data. Each one observation of duplicate genes has
been removed. See source section in this help for original data source.
Source
http://genome-www.stanford.edu/lung_cancer/adeno/
References
Garber, M. E. et al. (2001) "Diversity of gene expression in adenocarcinoma of the lung", Proceedings of the National Academy of Sciences, 98, 13784-13789.
Examples
## Reading the data
data(lung)
## Multiscale Bootstrap Resampling
lung.pv <- pvclust(lung, nboot=100)
## CAUTION: nboot=100 may be too small for actual use.
## We suggest nboot=1000 or larger.
## plot/print functions will be useful for diagnostics.
## Plot the result
plot(lung.pv, cex=0.8, cex.pv=0.7)
ask.bak <- par()$ask
par(ask=TRUE)
msfit 3
pvrect(lung.pv, alpha=0.9)
msplot(lung.pv, edges=c(51,62,68,71))
par(ask=ask.bak)
## Print a cluster with high p-value
lung.pp <- pvpick(lung.pv, alpha=0.9)
lung.pp$clusters[[2]]
## Print its edge number
lung.pp$edges[2]
## We recommend parallel computing for large dataset as this one
## Not run:
library(snow)
cl <- makeCluster(10, type="MPI")
lung.pv <- parPvclust(cl, lung, nboot=1000)
## End(Not run)
msfit Curve Fitting for Multiscale Bootstrap Resampling
Description
msfit performs curve fitting for multiscale bootstrap resampling. It generates an object of class
msfit. Several generic methods are available.
Usage
msfit(bp, r, nboot)
## S3 method for class 'msfit'
plot(x, curve=TRUE, main=NULL, sub=NULL, xlab=NULL, ylab=NULL, ...)
## S3 method for class 'msfit'
lines(x, col=2, lty=1, ...)
## S3 method for class 'msfit'
summary(object, digits=3, ...)
Arguments
bp numeric vector of bootstrap probability values.
r numeric vector of relative sample size of bootstrap samples defined as r = n′/n
for original sample size n and bootstrap sample size n′.
nboot numeric value (vector) of the number of bootstrap replications.
4 msfit
x object of class msfit.
curve logical. If TRUE, the fitted curve is drawn.
main, sub, xlab, ylab, col, lty
generic graphic parameters.
object object of class msfit.
digits integer indicating the precision to be used in rounding.
... other parameters to be used in the functions.
Details
function msfit performs the curve fitting for multiscale bootstrap resampling. In package pvclust
this function is only called from the function pvclust (or parPvclust), and may never be called
from users. However one can access a list of msfit objects by x$msfit, where x is an object of
class pvclust.
Value
msfit returns an object of class msfit. It contains the following objects:
p numeric vector of p-values. au is AU (Approximately Unbiased) p-value computed by multiscale bootstrap resampling, which is more accurate than BP value
(explained below) as unbiased p-value. bp is BP (Bootstrap Probability) value,
which is simple but tends to be unbiased when the absolute value of c (a value
in coef vector, explained below) is large.
se numeric vector of estimated standard errors of p-values.
coef numeric vector related to geometric aspects of hypotheses. v is signed distance
and c is curvature of the boundary.
df numeric value of the degree of freedom in curve fitting.
rss residual sum of squares.
pchi p-value of chi-square test based on asymptotic theory.
Author(s)
Ryota Suzuki
References
Shimodaira, H. (2004) "Approximately unbiased tests of regions using multistep-multiscale bootstrap resampling", Annals of Statistics, 32, 2616-2641.
Shimodaira, H. (2002) "An approximately unbiased test of phylogenetic tree selection", Systematic
Biology, 51, 492-508.
msplot 5
msplot Drawing the Results of Curve Fitting for Pvclust Object
Description
draws the results of curve fitting for pvclust object.
Usage
msplot(x, edges=NULL, ...)
Arguments
x object of class pvclust.
edges numeric vector of edge numbers to be plotted.
... other parameters to be used in the function.
Author(s)
Ryota Suzuki
See Also
plot.msfit
plot.pvclust Draws Dendrogram with P-values for Pvclust Object
Description
plot dendrogram for a pvclust object and add p-values for clusters.
Usage
## S3 method for class 'pvclust'
plot(x, print.pv=TRUE, print.num=TRUE, float=0.01,
col.pv=c(2,3,8), cex.pv=0.8, font.pv=NULL, col=NULL, cex=NULL,
font=NULL, lty=NULL, lwd=NULL, main=NULL, sub=NULL, xlab=NULL, ...)
## S3 method for class 'pvclust'
text(x, col=c(2,3,8), print.num=TRUE, float=0.01, cex=NULL, font=NULL, ...)
6 plot.pvclust
Arguments
x object of class pvclust, which is generated by function pvclust. See pvclust
for details.
print.pv logical flag to specify whether print p-values above the edges (clusters).
print.num logical flag to specify whether print edge numbers below clusters.
float numeric value to adjust the height of p-values from edges.
col.pv numeric vector of length three to specify the colors for p-values and edge numbers. From the beginning each value corresponds to the color of AU values, BP
values and edge numbers, respectively.
cex.pv numeric value which specifies the size of characters for p-values and edge numbers. See cex argument for par.
font.pv numeric value which specifies the font of characters for p-values and edge numbers. See font argument for par.
col, cex, font in text function, they correspond to col.pv, cex.pv and font.pv in plot function, respectively. In plot function they are used as generic graphic parameters.
lty, lwd, main, sub, xlab, ...
generic graphic parameters. See par for details.
Details
This function plots a dendrogram with p-values for given object of class pvclust. AU p-value
(printed in red color in default) is the abbreviation of "approximately unbiased" p-value, which
is calculated by multiscale bootstrap resampling. BP value (printed in green color in default) is
"bootstrap probability" value, which is less accurate than AU value as p-value. One can consider
that clusters (edges) with high AU values (e.g. 95%) are strongly supported by data.
Author(s)
Ryota Suzuki
References
Shimodaira, H. (2004) "Approximately unbiased tests of regions using multistep-multiscale bootstrap resampling", Annals of Statistics, 32, 2616-2641.
Shimodaira, H. (2002) "An approximately unbiased test of phylogenetic tree selection", Systematic
Biology, 51, 492-508.
See Also
text.pvclust
print.pvclust 7
print.pvclust Print Function for Pvclust Object
Description
print clustering method and distance measure used in hierarchical clustering, p-values and related
statistics for a pvclust object.
Usage
## S3 method for class 'pvclust'
print(x, which=NULL, digits=3, ...)
Arguments
x object of class pvclust.
which numeric vector which specifies the numbers of edges (clusters) of which the
values are printed. If NULL is given, it prints the values of all edges. The default
is NULL.
digits integer indicating the precision to be used in rounding.
... other parameters used in the function.
Value
this function prints p-values and some related statistics.
au AU (Approximately Unbiased) p-value, which is more accurate than BP value
as unbiased p-value. It is computed by multiscale bootstrap resampling.
bp BP (Bootstrap Probability) value, which is a simple statistic computed by bootstrap resampling. This value tends to be biased as p-value when the absolute
value of c (explained below) is large.
se.au, se.bp estimated standard errors for au and bp, respectively.
v, c values related to geometric aspects of hypotheses. v is signed distance and c is
curvature of the boundary.
pchi p-values of chi-square test based on asymptotic theory.
Author(s)
Ryota Suzuki
8 pvclust
pvclust Calculating P-values for Hierchical Clustering
Description
calculates p-values for hierarchical clustering via multiscale bootstrap resampling. Hierarchical
clustering is done for given data and p-values are computed for each of the clusters.
Usage
pvclust(data, method.hclust="average",
method.dist="correlation", use.cor="pairwise.complete.obs",
nboot=1000, parallel=FALSE, r=seq(.5,1.4,by=.1),
store=FALSE, weight=FALSE, iseed=NULL, quiet=FALSE)
parPvclust(cl=NULL, data, method.hclust="average",
method.dist="correlation", use.cor="pairwise.complete.obs",
nboot=1000, r=seq(.5,1.4,by=.1), store=FALSE, weight=FALSE,
init.rand=NULL, iseed=NULL, quiet=FALSE)
Arguments
data numeric data matrix or data frame.
method.hclust the agglomerative method used in hierarchical clustering. This should be (an abbreviation of) one of "average", "ward.D", "ward.D2", "single", "complete",
"mcquitty", "median" or "centroid". The default is "average". See method
argument in hclust.
method.dist the distance measure to be used. This should be a character string, or a function
which returns a dist object. A character string should be (an abbreviation of)
one of "correlation", "uncentered", "abscor" or those which are allowed
for method argument in dist function. The default is "correlation". See
details section in this help and method argument in dist.
use.cor character string which specifies the method for computing correlation with data
including missing values. This should be (an abbreviation of) one of "all.obs",
"complete.obs" or "pairwise.complete.obs". See the use argument in cor
function.
nboot the number of bootstrap replications. The default is 1000.
parallel switch for parallel computation. If FALSE the computation is done in nonparallel mode. If TRUE or a positive integer is supplied, parallel computation
is done with automatically generated PSOCK cluster. Use TRUE for default cluster size (parallel::detectCores() - 1), or specify the size by an integer. If
a cluster object is supplied the cluster is used for parallel computation. Note
that NULL is currently not allowed for using the default cluster.
pvclust 9
r numeric vector which specifies the relative sample sizes of bootstrap replications. For original sample size n and bootstrap sample size n′, this is defined as
r = n′/n.
store locical. If store=TRUE, all bootstrap replications are stored in the output object.
The default is FALSE.
cl a cluster object created by package parallel or snow. If NULL, use the registered default cluster.
weight logical. If weight=TRUE, resampling is made by weight vector instead of index
vector. Useful for large r value (r>10). Currently, available only for distance
"correlation" and "abscor".
init.rand logical. If init.rand=TRUE, random number generators are initialized. Use
iseed argument to achieve reproducible results. This argument is duplicated
and will be unavailable in the future.
iseed An integer. If non-NULL value is supplied random number generators are initialized. It is passed to set.seed or clusterSetRNGStream.
quiet logical. If TRUE it does not report the progress.
Details
Function pvclust conducts multiscale bootstrap resampling to calculate p-values for each cluster
in the result of hierarchical clustering. parPvclust is the parallel version of this procedure which
depends on package parallel for parallel computation.
For data expressed as (n× p) matrix or data frame, we assume that the data is n observations of p
objects, which are to be clustered. The i’th row vector corresponds to the i’th observation of these
objects and the j’th column vector corresponds to a sample of j’th object with size n.
There are several methods to measure the dissimilarities between objects. For data matrix X =
{xij}, "correlation" method takes
1−
∑n
i=1(xij − x̄j)(xik − x̄k)√∑n
i=1(xij − x̄j)2
√∑n
i=1(xik − x̄k)2
for dissimilarity between j’th and k’th object, where x̄j = 1n
∑n
i=1 xijandx̄k =
1
n ∑n
i=1 xik.
"uncentered" takes uncentered sample correlation
1−
∑n
i=1 xijxik√∑n
i=1 x
2
ij √∑n
i=1 x
2
ik and "abscor" takes the absolute value of sample correlation
1−
∣∣∣∣∣
∑n
i=1(xij − x̄j)(xik − x̄k)√∑n
i=1(xij − x̄j)2
√∑n
i=1(xik − x̄k)2
∣∣∣∣∣.
Value
hclust hierarchical clustering for original data generated by function hclust. See
hclust for details.
10 pvclust
edges data frame object which contains p-values and supporting informations such as
standard errors.
count data frame object which contains primitive information about the result of multiscale bootstrap resampling.
msfit list whose elements are results of curve fitting for multiscale bootstrap resampling, of class msfit. See msfit for details.
nboot numeric vector of number of bootstrap replications.
r numeric vector of the relative sample size for bootstrap replications.
store list contains bootstrap replications if store=TRUE was given for function pvclust
or parPvclust.
Author(s)
Ryota Suzuki
References
Suzuki, R. and Shimodaira, H. (2006) "Pvclust: an R package for assessing the uncertainty in
hierarchical clustering", Bioinformatics, 22 (12): 1540-1542.
Shimodaira, H. (2004) "Approximately unbiased tests of regions using multistep-multiscale bootstrap resampling", Annals of Statistics, 32, 2616-2641.
Shimodaira, H. (2002) "An approximately unbiased test of phylogenetic tree selection", Systematic
Biology, 51, 492-508.
Suzuki, R. and Shimodaira, H. (2004) "An application of multiscale bootstrap resampling to hierarchical clustering of microarray data: How accurate are these clusters?", The Fifteenth International
Conference on Genome Informatics 2004, P034.
http://www.sigmath.es.osaka-u.ac.jp/shimo-lab/prog/pvclust/
See Also
lines.pvclust, print.pvclust, msfit, plot.pvclust, text.pvclust, pvrect and pvpick.
Examples
### example using Boston data in package MASS
data(Boston, package = "MASS")
## multiscale bootstrap resampling (non-parallel)
boston.pv <- pvclust(Boston, nboot=100, parallel=FALSE)
## CAUTION: nboot=100 may be too small for actual use.
## We suggest nboot=1000 or larger.
## plot/print functions will be useful for diagnostics.
## plot dendrogram with p-values
plot(boston.pv)
ask.bak <- par()$ask
pvpick 11
par(ask=TRUE)
## highlight clusters with high au p-values
pvrect(boston.pv)
## print the result of multiscale bootstrap resampling
print(boston.pv, digits=3)
## plot diagnostic for curve fitting
msplot(boston.pv, edges=c(2,4,6,7))
par(ask=ask.bak)
## print clusters with high p-values
boston.pp <- pvpick(boston.pv)
boston.pp
### Using a custom distance measure
## Define a distance function which returns an object of class "dist".
## The function must have only one argument "x" (data matrix or data.frame).
cosine <- function(x) {
x <- as.matrix(x)
y <- t(x) %*% x
res <- 1 - y / (sqrt(diag(y)) %*% t(sqrt(diag(y))))
res <- as.dist(res)
attr(res, "method") <- "cosine"
return(res)
}
result <- pvclust(Boston, method.dist=cosine, nboot=100)
plot(result)
## Not run:
### parallel computation
result.par <- pvclust(Boston, nboot=1000, parallel=TRUE)
plot(result.par)
## End(Not run)
pvpick Find Clusters with High/Low P-values
Description
find clusters with relatively high/low p-values. pvrect and lines (S3 method for class pvclust)
highlight such clusters in existing plot, and pvpick returns a list of such clusters.
12 pvpick
Usage
pvpick(x, alpha=0.95, pv="au", type="geq", max.only=TRUE)
pvrect(x, alpha=0.95, pv="au", type="geq", max.only=TRUE, border=2, ...)
## S3 method for class 'pvclust'
lines(x, alpha=0.95, pv="au", type="geq", col=2, lwd=2, ...)
Arguments
x object of class pvclust.
alpha threshold value for p-values.
pv character string which specifies the p-value to be used. It should be either of
"au" or "bp", corresponding to AU p-value or BP value, respectively. See
plot.pvclust for details.
type one of "geq", "leq", "gt" or "lt". If "geq" is specified, clusters with p-value
greater than or equals the threshold given by "alpha" are returned or displayed.
Likewise "leq" stands for lower than or equals, "gt" for greater than and "lt"
for lower than the threshold value. The default is "geq".
max.only logical. If some of clusters with high/low p-values have inclusion relation, only
the largest cluster is returned (or displayed) when max.only=TRUE.
border numeric value which specifies the color of borders of rectangles.
col numeric value which specifies the color of lines.
lwd numeric value which specifies the width of lines.
... other graphic parameters to be used.
Value
pvpick returns a list which contains the following values.
clusters a list of character string vectors. Each vector corresponds to the names of objects
in each cluster.
edges numeric vector of edge numbers. The i’th element (number) corresponds to the
i’th name vector in clusters.
Author(s)
Ryota Suzuki
seplot 13
seplot Diagnostic Plot for Standard Error of p-value
Description
draws diagnostic plot for standard error of p-value for pvclust object.
Usage
seplot(object, type=c("au", "bp"), identify=FALSE, main=NULL,
xlab=NULL, ylab=NULL, ...)
Arguments
object object of class pvclust.
type the type of p-value to be plotted, one of "au" or "bp".
identify logical. If TRUE, edge numbers can be identified interactively. See identify for
basic usage.
main, xlab, ylab
generic graphic parameters. See par for details.
... other graphical parameters to be passed to generic plot or identify function.
Author(s)
Ryota Suzuki
Index
∗Topic aplot
pvpick, 11
∗Topic cluster
pvclust, 8
∗Topic datasets
lung, 2
∗Topic hplot
msplot, 5
plot.pvclust, 5
seplot, 13
∗Topic htest
msfit, 3
∗Topic print
print.pvclust, 7
cor, 8
dist, 8
hclust, 8, 9
identify, 13
lines.msfit (msfit), 3
lines.pvclust, 10
lines.pvclust (pvpick), 11
lung, 2
msfit, 3, 10
msplot, 5
par, 6, 13
parPvclust (pvclust), 8
plot.msfit, 5
plot.msfit (msfit), 3
plot.pvclust, 5, 10
print.pvclust, 7, 10
pvclust, 6, 8
pvpick, 10, 11
pvrect, 10
pvrect (pvpick), 11
seplot, 13
summary.msfit (msfit), 3
text.pvclust, 6, 10
text.pvclust (plot.pvclust), 5
14

PDF Document reader online

This website is focused on providing document in readable format, online without need to install any type of software on your computer. If you are using thin client, or are not allowed to install document reader of particular type, this application may come in hand for you. Simply upload your document, and Docureader.top will transform it into readable format in a few seconds. Why choose Docureader.top?

  1. Unlimited sharing - you can upload document of any size. If we are able to convert it into readable format, you have it here - saved for later or immediate reading
  2. Cross-platform - no compromised when reading your document. We support most of modern browers without the need of installing any of external plugins. If your device can oper a browser - then you can read any document on it
  3. Simple uploading - no need to register. Just enter your email, title of document and select the file, we do the rest. Once the document is ready for you, you will receive automatic email from us.

Previous 10

Next 10