Test Set Generator

Search Mode:
Query	PDB Ensemble	Key word

The Query mode allows the user to select structural and quality filters to generate a test set at any given level of residual sequence and structural similarity using the CATH structural classification scheme. CATH defines several levels of structural as well as common sequence similarity thresholds (e.g. less than 35% sequence identity). Additionally, the server implements several quality checks for the structures to be included, e.g. X-ray resolution and R-free cutoffs. The data can be visualized either interactively, for smaller sets, or simply downloaded.

The PDB Ensemble mode can be used to seed the structural search form a limited number of PDB codes of proteins sharing the desired structural and/or sequence features. The server will then present a list of CATH codes from which to choose the desired set in analogy to the Query mode. This can be useful if the intention is to extend a previously published test set.

The Key word mode initiates a structural search from key words contained in the header and compound records of all PDB structures. A list of matching proteins is then presented and can be manipulated in analogy to the Query mode. This can be useful if the user has no specific idea about the PDB codes of relevant proteins or their structural classification.