SHREC 2007-Protein Challenge
| Start | Data | Evaluation | Register | Misc | Results |
Welcome to the Freiburg track on proteins for the SHREC(SHape REtrieval Contest) 2007 competition!
This competition is organized by the AIM@SHAPE network.
In our track we would like to evaluate the effectiveness of 3D features on protein 3D models. Since the classification of protein structures is mainly done by biologists, chemists and crystallographers, we would like to invite the shape retrieval community to test their skills on this task.
 |
|
| Motivation: | Proteins are linear sequences of amino acids which fold into three dimensional structures.
Throughout evolution the amino acid composition can change, but the three dimensional
structure of the protein stays conserved. The three dimensional structure of a protein is
closely linked to its function. So, by finding similar three dimensional protein structures,
their function and evolutionary linkage can be determined. Molecular biologists are often interested in getting a survey of the objects in a biomolecular database making classification one of their basic tasks: To which of the recognized classes in the database does a new molecule belong? Several classification schemata such as SCOP [5], CATH [6] and DALI/FSSP [2] are available in the Internet. When a new object is inserted into the database the supervision by experts that are very experienced and have a deep knowledge in the domain of molecular biology is necessary in most cases. An efficient classification algorithm is desired that can speed up the classification process by acting as a fast filter for further investigation. While SCOP and CATH require classification by human experts, a fully automatic classification is available from the FSSP database (Families of Structurally Similar Proteins), generated by the DALI (Distance matrix ALIgnment) system. The evaluation of a pair of proteins is very expensive, since query processing for a single molecule against the entire FSSP database currently takes an overnight run. |
| The Task : | The task of this competition is to classify protein domains to one of the SCOP folds. The
participants will train their feature extraction algorithms on the provided data set. Then,
one day before the end of the competition, the participants will be provided with a set of 30
unknown protein domains. The query files will contain all atoms of the protein domain and
their 3D coordinates. The task is then to assign the query protein domains to SCOP folds. Since the entire SCOP database is divided into more than 970 folds, we will limit the task to assigning the unknown protein domains to one of the 27 folds provided in the data section. |
| Time Schedule : |
April 1st: All details published |
| Submission: | Send via e-mail |
This homepage was created by Maja Temerinac for the SHREC 2007 Protein Challenge.
For more information, please contact:
Maja Temerinac temerina(at)informatik.uni-freiburg.de or
Marco Reisert reisert(at)informatik.uni-freiburg.de