I'm comparing two ways of creating heatmaps with dendrograms in R, one with made4's heatplot (from bioconductor) and one with gplots of heatmap.2. The appropriate results depend on the analysis but I'm trying to understand why the defaults are so different, and how to get both functions to give the same result (or highly similar result) so that I understand all the 'blackbox' parameters that go into this.
Bioconductor version: Release (3.12) The package offer different classifiers based on comparisons of pair of features (TSP), using various decision rules (e.g., majority wins principle). Author: Bahman Afsari, Luigi Marchionni, Wikum Dinalankara. Gplots documentation built on Nov. 28, 2020, 5:07 p.m. CRAN packages Bioconductor packages R-Forge packages GitHub packages. We want your feedback!
This is the example data and packages:
Clustering the data with heatmap.2 gives:
Using heatplot gives:
very different results and scalings initially. heatplot results look more reasonable in this case so I'd like to understand what parameters to feed into heatmap.2 to get it to do the same, since heatmap.2 has other advantages/features I'd like to use and because I want to understand the missing ingredients.
heatplot uses average linkage with correlation distance so we can feed that into heatmap.2 to ensure similar clusterings are used (based on: https://stat.ethz.ch/pipermail/bioconductor/2010-August/034757.html)
resulting in:
this makes the row-side dendrograms look more similar but the columns are still different and so are the scales. It appears that heatplot scales the columns somehow by default that heatmap.2 doesn't do that by default. If I add a row-scaling to heatmap.2, I get:
which still isn't identical but is closer. How can I reproduce heatplot's results with heatmap.2? What are the differences?
edit2: it seems like a key difference is that heatplot rescales the data with both rows and columns, using:
this is what I'm trying to import to my call to heatmap.2. The reason I like it is because it makes the contrasts larger between the low and high values, whereas just passing zlim to heatmap.2 gets simply ignored. How can I use this 'dual scaling' while preserving the clustering along the columns? All I want is the increased contrast you get with:
heatplot(..., dualScale=TRUE, scale='none')
compared with the low contrast you get with:
heatplot(..., dualScale=FALSE, scale='row')
any ideas on this?
Plots Bioconductor
Using Bioconductor
The current release of Bioconductor is version3.12; it works with R version4.0.3. Users of older R andBioconductor must update their installation to take advantageof new features and to access packages that have been added toBioconductor since the last release.
The development version of Bioconductor is version3.13; it works with R version4.1.0. More recent ‘devel’versions of R (if available) will be supported during the nextBioconductor release cycle.
Install the latest release of R, then get the latest version ofBioconductor by starting R and entering the commands
It may be possible to change the Bioconductor version of an existinginstallation; see the ‘Changing version’ section of the BiocManagervignette.
Details, including instructions toinstall additional packages and toupdate,find, andtroubleshoot are providedbelow. A devel version ofBioconductor is available. There are goodreasons for using BiocManager::install() formanaging Bioconductor resources.
Install R
- Download the most recent version of R. The R FAQs and the RInstallation and Administration Manual contain detailed instructionsfor installing R on various platforms (Linux, OS X, and Windows beingthe main ones).
Start the R program; on Windows and OS X, this will usually meandouble-clicking on the R application, on UNIX-like systems, type“R” at a shell prompt.
As a first step with R, start the R help browser by typing
help.start()in the R command window. For help on anyfunction, e.g. the “mean” function, type? mean.
[ Back to top ]
Install Bioconductor Packages
To install core packages, type the following in an R command window:
Install specific packages, e.g., “GenomicFeatures” and “AnnotationDbi”, with
The install() function (in the BiocManager package) has arguments that changeits default behavior; type ?install for further help.
For a more detailed explanation on using BiocManager and its advanced usage,such as version switching, please refer to theBiocManager vignette.
[ Back to top ]
Find Bioconductor Packages
Visit the software package listto discover available packages.
To search through available packages programmatically, use the following:
For example, using a “^org” search pattern will show all of the availableorganism annotation packages.
[ Back to top ]
Update Installed Bioconductor Packages
Bioconductor packages, especially those in the development branch, areupdated fairly regularly. To identify packages requiring update withinyour version of Bioconductor, start a new session of R and enter
Use the argument ask=FALSE to update old packages without beingprompted. Read the help page for ?install for additional details.
Upgrading installed Bioconductor packages
Some versions of R support more than one version of Bioconductor. Touse the latest version of Bioconductor for your version of R, enter
Remember that more recent versions of Bioconductor may be available if yourversion of R is out-of-date.
For more details on Bioconductor approaches to versioning, seethe advanced sectionin the vignette and version numbering in the developer reference section.
Recompiling installed Bioconductor packages
Rarely, underlying changes in the operating system require ALLinstalled packages to be recompiled for source (C or Fortran)compatibility. One way to address this might be to start a new Rsession and enter
As this will reinstall all currently installed packages, it likelyinvolves a significant amount of network bandwidth and compilationtime. All packages are implicitly updated, and the cumulative effectmight introduce wrinkles that disrupt your work flow. It also requiresthat you have the necessary compilers installed.
[ Back to top ]
Gplots Bioconductor
Troubleshoot Package Installations
Use the commands
to flag packages that are either out-of-date or too new for yourversion of Bioconductor. The output suggests ways to solve identifiedproblems, and the help page ?valid lists arguments influencingthe behavior of the function.
Troubleshoot BiocManager
One likely reason for BiocManager not working on your system couldbe that your version of R is too old for BiocManager. In orderavoid this issue, please ensure that you have the latest version of Rinstalled in your system. BiocManager supports R versions from 3.5.0and above.
[ Back to top ]
Why use BiocManager::install()?
BiocManager::install() is the recommended way to install Bioconductorpackages. There are several reasons for preferring this to the‘standard’ way in which R pacakges are installed viainstall.packages().
Bioconductor has a repository and release schedule that differs from R(Bioconductor has a ‘devel’ branch to which new packages and updatesare introduced, and a stable ‘release’ branch emitted once every 6months to which bug fixes but not new features are introduced).
A consequence of the mismatch between R and Bioconductor releaseschedules is that the Bioconductor version identified byinstall.packages() is sometimes not the most recent ‘release’available. For instance, an R minor version may be introduced somemonths before the next Bioc release. After the Bioc release the usersof the R minor version will be pointed to an out-of-date version ofBioconductor.
A consequence of the distinct ‘devel’ branch is thatinstall.packages() sometimes points only to the ‘release’repository, whereas Bioconductor developers and users wantingleading-edge features wish to access the Bioconductor ‘devel’repository. For instance, the Bioconductor 3.0 release is availablefor R.3.1.x, so Bioconductor developers and leading-edge users need tobe able to install the devel version of Bioconductor packages into thesame version (though perhaps different instance or at least librarylocation) of R that supports version 2.14 of Bioconductor.
An indirect consequence of Bioconductor’s structured release is thatpackages generally have more extensive dependencies with one another,both explicitly via the usual package mechanisms and implicitlybecause the repository, release structure, and Bioconductor communityinteractions favor re-use of data representations and analysisconcepts across packages. There is thus a higher premium on knowingthat packages are from the same release, and that all packages arecurrent within the release.
The BiocManager package serves as the primary way to ensure thatthe appropriate Bioconductor installation is used with respectto the version of R in use regardless of the R and Bioconductorrelease cycles.
The install() function is provided by BiocManager. This is awrapper around install.packages, but with the repository chosenaccording to the version of Bioconductor in use, rather than to theversion relevant at the time of the release of R.
install() also nudges users to remain current within a release, bydefault checking for out-of-date packages and asking if the user wouldlike to update
The BiocManager package provides facilities for switching to the‘devel’ version of Bioconductor
(at some points in the R / Bioconductor release cycle use of ‘devel’requires use of a different version of R itself, in which case theattempt to install devel fails with an appropriate message).
The BiocManager package also provides valid() to test that theinstalled packages are not a hodgepodge from different Bioconductorreleases (the ‘too new’ packages have been installed from sourcerather than a repository; regular users would seldom have these).
For users who spend a lot of time in Bioconductor, the featuresoutlined above become increasingly important and install() is muchpreferred to install.packages().
[ Back to top ]
Pre-configured Bioconductor
Bioconductor is also available as a set ofAmazon Machine Images (AMIs) andDocker images.
Legacy and Older R Versions
It is always recommended to update to the most current version of R andBioconductor. If this is not possible and R < 3.5.0 , please use the followingfor installing Bioconductor packages
To install core packages, type the following in an R command window:
Install specific packages, e.g., “GenomicFeatures” and “AnnotationDbi”, with
[ Back to top ]