ProCope - Protein Complex Prediction and Evaluation
ProCope is a software package which provides easy access to
different methods used for the prediction and evaluation of protein
complexes from purification data experiments.
Large scale protein purification experiments provide a vast amount of
indirect protein-protein interaction information. As these data are
inherently noisy and incomplete, sophisticated methods are required to
predict actual protein complexes existing and operating in-vivo. Since the
publication of two large-scale purification datasets for the yeast Saccharomyces cerevisiae
in 2006 (Gavin et al. and Krogan et al.) different methods
for the processing of these datasets have been developed. These include
methods for the derivation of PPI score networks from the data, the
clustering of these networks to predict actual complexes and the
evaluation of the predictions using experimental data.
ProCope implements many of the methods described in the literature
along with further useful methods for PPI score network manipluation,
complex set manipulation, data integration with Petri nets and result
visualization. All functionality can be accessed using a graphical user interface,
a collection of command
line tools and an application
programming interface
(API) to use all these features in
your own computer programs. The whole software package is written in
Java and requires the
Java Runtime Environment 5.0 or
higher.
Getting started
Installation guide
If you quickly want to start using ProCope check out the first steps guide of the graphical user interface.
Parts of the package
Detailed documentations of the three differents parts of the package
are linked here and above in the navigation panel.
Methods and Materials
Here you find information about the datasets and methods implemented in the package
and references to the original literature. Note the navigation panel on
the left which gives you quick access to all parts of this
documentation.
- Datasets - describes the basic
dataset objects used in the package, namely protein complex sets,
purification data sets and protein score networks
- Score network generation - gives an overview over the scoring methods used to derive PPI confidence values from a set of purification data
- Complex set prediction - basic concepts of complex set prediction using the clustering of protein score networks and the evaluation of the results
- Complex set evaluation - methods to evaluate predicted complexes using experimental data
- Network evaluation - the quality of PPI score networks can directly be evaluated with experiment data
- Petri nets - ProCope can
integrate information if various datasets into single Petri nets and
export the result to different file formats (currently ToPNet and XGMML)
Additional information and features
- Networks - basic information about the protein networks in ProCope
- File formats - ProCope awaits certain file formats for the different types of data
- Name mappings - for many proteins there are different identifiers from different databases, use name mappings to solve this problem
- Annotations and filtering
- proteins and edges in protein networks can be annotated with
arbitrary key/value pairs, these annotations can be used for filtering
- User defined classes - GUI and command line interface can be extended by user-defined clustering implementations and score calculators
- Interologs - ProCope contains some methods to calculate and score interologs
- External libraries - External software libraries used by ProCope