This page is written for an audience of scientists new to Rosetta: perhaps a first year graduate student, or young postdoc, who has received/started a project that needs "some computer modeling". In other words, an individual coming to Rosetta from a cold start. Is Rosetta a good tool for the modeling you need to do? If so, how do you go about getting and using Rosetta? If you are already comfortable with the concepts, feel free to skip ahead.
Rosetta is a very large software suite for macromolecular modeling. By software suite, we mean that it is a large collection of computer code (mostly in C++, some in Python, a little in other languages), but it is not a single monolithic program. By macromolecular modeling, we mean the process of evaluating and ranking the physical plausibility of different structures of biological macromolecules (usually protein, but nucleic acids and ligands are significantly supported and support for implicit lipid membranes is growing). Generally, a user will choose some specific protocol within Rosetta and provide that protocol with inputs for A) what structure to work on, and B) what options within the protocol are appropriate for the user's needs.
Doing macromolecular modeling well—doing good science—requires careful consideration of your inputs, how the modeling is performed, and analysis of your outputs. Rosetta itself can be operated as a "black box", but you are doing yourself and your science a disservice if you use it that way.
1) Inputs to Rosetta
The major input to Rosetta is the input structure. Generally, if you have a high-resolution structure(s) (better than 2 Å) of your molecule, it can be used with Rosetta with few changes. If you have a poorer-resolution structure, an NMR structure, a homology model, or no structure at all, then your task will be much harder. You should still prepare your structure for modeling, but be aware that modeling is less efficient and effective when starting from poor quality structures and interpreting results is more challenging.
2) Choosing the Rosetta protocol
3) Computational resources
Rosetta software, as a whole, is written to run on supercomputers, but can be run on many different scales. Most applications can give "dry runs" for testing on any computer. A few applications can run on laptop computers. There are applications than can run on lab-scale powerful computers (12-30 core range). Most applications assume you have access to tens of thousands of hours of computer time to accumulate enough results to answer your question. The later sections of this document describe installing or using Rosetta at those different scales.
For commercial use, see the section below and the Rosetta Servers page.
ROSIE is a server that offers several Rosetta applications through a simple web interface. It is perfect for use by those new to Rosetta. Despite ROSIE's variety it offers only a slice of Rosetta's full functionality.
ROBETTA (Robot-Rosetta) is a server that provides ab initio folding and structure prediction, as well as fragment selection for local runs of Rosetta.
For academic or commercial users, you can request a license. These licenses are free for academic users. Once you have a license, you can download the code here. Note that the download comes as a tar archive. We don't distribute executables/binaries for most purposes, we distribute source code. As a consequence, you will need to compile the code before use.
After downloading Rosetta, go through these detailed tutorials to get started. Once you are done with the introductory tutorials, you should be able to do most common molecular modeling tasks on Rosetta.
Local installation implies that one will be using Rosetta through a command line interface (or terminal). For local users, you are unlikely to want to install Rosetta to the entire system. Rosetta is quite happy to be compiled and installed by regular users without administrative rights — this is how the developers use it. You may need administrative rights to install dependencies.
First untar/uncompress your downloaded copy of the code (
tar -xvzf Rosetta[releasenumber].tar.gz).
Next, navigate to the
The basic compilation command is
./scons.py -j<number_of_processors_to_use> mode=release bin.
Replace with a number one processor fewer than your computer has.
Expect compilation to take a while (hours on one processor).
See our extensive build documentation for further instructions and troubleshooting.
Unfortunately, we are not currently able to support the whole of Rosetta on Windows. There are few free, easy-to-use C++ compilers available for Windows, and they use slightly different C++ standards. Dual booting or virtual machines running Linux/MacOS are options. We cannot help set up Windows/Linux dual boots, but we can help with Rosetta on the Linux partition. A subset of Rosetta that is required by PyRosetta, a Python-based interface to Rosetta, is supported on Windows and made available on http://www.pyrosetta.org.
If you'll be running Rosetta on a scientific computation cluster, there may already be a version of Rosetta installed for general usage. Talk to your cluster administrator to see if there is a centrally-provided version available for your use.
If your cluster doesn't have Rosetta already installed, or you wish to use a different version than the centrally installed one, don't worry — Rosetta is designed to be compiled and installed by regular users without administrative rights. As long as commonly available compilation tools are available for your use, you should be able to build and run Rosetta in your user directory without cluster administrator involvement. Just treat it like a local install into userspace.
Developers can update these doc pages through our interactive wiki [note: Github login required].