OverviewΒΆ

loopdb_builder is an application to build a database of loops in known structures, that can subsequently be used in Spruce to modeling missing loops.

Input for the application is a directory of PDB structures. An existing database can be appended to as new structures become available. OpenEye ships a pre-built database, which can be added to with internal/propriatary structures.

Loops are extracted from the structures in a sliding window producing loops of length 4, to the maximum length of a loop specified (default 22). The extracted loops are deduplicated so a loop with a identical sequences, rough distance between the anchor atoms, and similar backbone shape are only included once. The option to exclude “loops” with a regular secondary structure content (helix,strand) of >75% is on by default but can be turned off. The crop length removes N number of residues at the different termini of the protein structure before processing, knowing that the terminal residues are often flexible and poorly resolved in crystal structures, the default value is 1.