Font size: Increase font size Decrease font size Switch style sheet from default to highcontrast and back

The ViennaRNA Package

The ViennaRNA Package consists of a C code library and several stand-alone programs for the prediction and comparison of RNA secondary structures.

RNA secondary structure prediction through energy minimization is the most used function in the package. We provide three kinds of dynamic programming algorithms for structure prediction: the minimum free energy algorithm of (Zuker & Stiegler 1981) which yields a single optimal structure, the partition function algorithm of (McCaskill 1990) which calculates base pair probabilities in the thermodynamic ensemble, and the suboptimal folding algorithm of (Wuchty 1999) which generates all suboptimal structures within a given energy range of the optimal energy. For secondary structure comparison, the package contains several measures of distance (dissimilarities) using either string alignment or tree-editing (Shapiro & Zhang 1990). Finally, we provide an algorithm to design sequences with a predefined structure (inverse folding).

In case you are using our software for your publications you may want to cite:

Lorenz, Ronny and Bernhart, Stephan H. and Höner zu Siederdissen, Christian and Tafer, Hakim and Flamm, Christoph and Stadler, Peter F. and Hofacker, Ivo L.
ViennaRNA Package 2.0
Algorithms for Molecular Biology, 6:1 26, 2011, doi:10.1186/1748-7188-6-26


  • 01 - 25 - 2016

    Version 2.2 is out! After almost a year without a new release, we are happy to announce many new features. This version officially introduces (generic) hard- and soft-constraints for many of the folding algorithms. Thus, chemical probing constraints, such as derived from SHAPE experiments, can be easily incorporated into RNAfold, RNAalifold, and RNAsubopt. Furthermore, RNAfold and the RNAlib interface allow for a simple way to incorporate ligand binding to specific hairpin- or interior-loop motifs. This version also introduces the new v3.0 API of the RNAlib C-library, that will eventually replace the current interface in the future.
    See the Changelog for version 2.2.0 for a complete list of new features and bugfixes.

  • 02 - 05 - 2015

    Version 2.1.9 is a major bugfix release that changes the way how the ViennaRNA Package handles dangling end and terminal mismatch contributions for exterior-, and multibranch loops. We strongly recommend upgrading your installation to this or a newer version to obtain predictions that are better comparable to RNAstructure or UNAFold.
    Please see the Changelog for version 2.1.9 for further details on the actual changes to the underlying energy parameters.

  • 04 - 13 - 2014

    For a long time, Mac OS X users were not able to correctly build the Perl/Python interface of the ViennaRNA Package. Starting with v2.1.7, this limitation has been removed, and the interface should compile and work as expected. Please see the Install Notes for Mac OS X users for further details.

  • 01 - 25 - 2013

    Since ViennaRNA Package Version 2.1.0 we have enabled G-Quadruplex prediction support into RNAfold, RNAcofold, RNALfold, RNAalifold, RNAeval and RNAplot. See the changelog for details.

  • 05 - 21 - 2012

    A preliminary version of the ViennaRNA Package implementing RNA/DNA hybrid support can be found here

  • 08 - 01 - 2011

    With the new release of version 2, we introduce the most recent nearest neighbor energy model for all free energy calculations. Additionally, most of the stand-alone programs included are now able to read FASTA formatted input data. This makes conversion of you data files, as necessary in previous releases, obsolete!

Get the latest ViennaRNA Package

The latest stable release is Version 2.3.3 from January, 24th 2017

Get the source code and compile it yourself Compile from Source Code

Installing from sourcecode is the recommended way to get most out of the ViennaRNA Package. For best portability the ViennaRNA package uses the GNU tools autoconf, automake, and libtool and can thus be compiled and installed on almost every computer platform. See the INSTALL instructions for details.

For those interested in contributing to the development of the ViennaRNA Package by sending patches that add functionalities or fix bugs, we provide access to the master branch of the ViennaRNA Package repository at github.

Install precompiled binary package Install precompiled Binary Package

For those unable to compile the ViennaRNA Package, or who simply don't want to, we also provide precompiled binary packages for several Linux distributions and installer executables for Windows and MacOS X. To comply with the packaging standards of the corresponding Linux distributions we split the ViennaRNA Package into

usually the executable programs and utility scripts
Development Files...
the RNAlib C-library and the corresponding header files
Perl Bindings...
The RNA module to access RNAlib C-library functions from within Perl scripts
Python Bindings...
The RNA class to access RNAlib C-library functions from within Python scripts

Simply select your Linux distribution or operating system by clicking on the corresponding icon below, and download the binary packages that best suite your needs from the corresponding list.
Ubuntu PPA Packages
The most convenient way to keep up-to-date with the latest ViennaRNA Packge is to use Ubuntu PPA:
sudo apt-add-repository ppa:j-4/vienna-rna
sudo apt-get update
sudo apt-get install vienna-rna
Note: The Windows installer will not change the PATH variable of your system. In order to omit the path to the binaries while using the programs of the ViennaRNA Package, please add the path of your ViennaRNA Package installation directory to your PATH variable manually.

Build your own package Build your own Binary Package from SourceCode

If you don't find a precompiled package for your linux distribution, or in case you just want to create a binary package by yourself, have a look into the packaging/ directory of our source tarball. It includes a list of some configuration files that we use for package generation:

First, for RPM based distributions, we use this SPEC file, originally written for Fedora, but meanwhile extended to work for at least the Linux distributions we provide binary packages ourselves.

Second, our binary package for ArchLinux was generated using this PKGBUILD file, which is also part of the viennarna source package available through the ArchLinux User Repository (AUR).

Third, the configuration files for our binary debian packages are included in the source tarball within the packaging/debian directory.

Finally, for our MacOS X package, we make use of the pkgbuild, productbuild, and hditool commandline tools. The corresponding configuration files are included in our tarball in the directory packaging/macosx. For simplicity, you can instruct the autotools toolchain to automatically build the installer disk image by adding the following configure switch:

./configure --enable-macosx-installer
This creates a file packaging/macosx/ViennaRNA-VERSION-MacOSX.dmg, where VERSION is the current version number.

Our binary packages in the above section are mainly build using the openSUSE Build Service, while the Windows Installer is created using the MinGW cross-compiler and the NSIS install system. The MacOS X installer is build on a MacBook Pro.

Performance Analysis

To investigate the impact of the new energy parameters used in ViennaRNA Package 2 we did a quite extensive performance analysis.

Update: With the release of ViennaRNA Package 2.1.9 we re-ran all performance benchmarks. Now, except for the runtime analysis, the charts and data compare RNAfold 1.8.5, RNAfold 2.0, UNAFold 3.8, RNAstructure 5.7, and, additionally, RNAfold 2.1.9.


We have measured the Performance of ViennaRNA Packge 2 by comparing Minimum Free Energy (MFE) predictions of the RNAfold program to

  1. RNAfold 1.8.5
  2. RNAfold 2.1.8
  3. UNAfold 3.8
  4. RNAstructure 5.7
both in terms of prediction accuracy as well as in computation speed, i.e. runtime.

Please note, that for this performance analysis we only compared thermodynamics-based approaches that rely on the Nearest neighbor energy model and solve the problem by Zuker's dynamic programming algorithm!
Benchmarks for other approaches in RNA structure prediction can be found elsewhere in literature or in the web.

Runtime analysis

Computational speed was measured using a dataset of randomly generated sequences with fixed lengths of

  • 100 nt (100 samples)
  • 500 nt (100 samples)
  • 1000 nt (100 samples)
  • 2500 nt (20 samples)
  • 5000 nt (16 samples)
  • 10000 nt (16 samples)
Measurements were taken on an Intel Core2 6600 CPU running at 2.4GHz.

Computation times (arithmetic mean)

Unfortunaltey, RNAstructure 5.7 was not able to predict an MFE structure for the 10000nt samples in a relatively small time frame and thus was omitted in the particular test.

As visible in the computation times graph above, we observe virtually no difference in the runtimes of RNAfold 1.8.5 and RNAfold 2.0. This is also true for the memory consumption (data not shown).

Update: Runtimes for RNAfold 2.1.9 does not differ substantially to that of RNAfold 2.0.

Although all programs in the test have the same asymptotic runtime complexity, the computation time analysis of RNAfold compares quite favorably to that of the competing implementations.

Prediction accuracy

We calculated the following four performance measures to assess the prediction accuracy:

  1. Sensitivity
  2. Positive Predictive Value (PPV)
  3. Matthews correlation coefficient (MCC)
  4. F-measure

The test set was based on a set comprising 1919 non-multimer sequence/structure pairs taken from the RNAstrand database (all without pseudoknots in the reference structure). Both versions of RNAfold were run with -d2 option whereas UNAfold and RNAstructure were run with default options.

In the table below, the resulting arithmetic mean of each performance measure is shown. Furthermore, we did a bootstrapping analysis with 1000 iterations to estimate the 95% confidence intervals for the predicted measures.

Program Sensitivity PPV MCC F-measure
RNAfold 2.1.9 0.742 [0.732,0.752] 0.795 [0.784,0.807] 0.767 [0.756,0.777] 0.765 [0.754,0.776]
RNAfold 2.1.8 0.740 [0.729,0.750] 0.792 [0.782,0.804] 0.764 [0.754,0.775] 0.762 [0.752,0.773]
RNAfold 1.8.5 0.711 [0.701,0.722] 0.773 [0.762,0.784] 0.740 [0.729,0.750] 0.737 [0.727,0.748]
UNAfold 3.8 0.693 [0.683,0.704] 0.767 [0.756,0.779] 0.727 [0.717,0.738] 0.725 [0.714,0.735]
RNAstructure 5.7 0.716 [0.707,0.727] 0.781 [0.771,0.792] 0.746 [0.737,0.757] 0.744 [0.736,0.753]

The cumulative distribution of the MCC shows that RNAfold 2.0 (represented by RNAfold 2.1.8, and RNAfold 2.1.9) outperformes the other programs on the test dataset: more of its predictions are within the region of higher performance values.

However, a detailed look at the performance among different RNA classes in our test set reveals that it differs widely. No single implementation tested provides consistent superiority of results.

Other Packages

Here we list all versions of the ViennaRNA Package that contain special features which have not been included into the main release yet:

Older Versions

There should rarely be a good reason to use any but the latest version of our software. However if you want to look up the old bugs, here's a list with most of the older releases for download.

Just click on the version numbers below to display the corresponding ViennaRNA Package releases.

Comments and Bug Reports

If in doubt our program is right, nature is at fault.
Comments and bug reports should be sent to