-
Notifications
You must be signed in to change notification settings - Fork 0
Local resources
StructMAn can make use of locally stored databases and use programs that require an user-specific license.
The functionalities of StructMan would be unthinkable without the structural information stored in the PDB. While StructMAn is capable of retrieving all necessary information directly online from the PDB, it is beneficial for the computation time and the network load to use a local instance of the PDB. This instance has to be stored in structman/resources/. It can be automatically downloaded there using:
structman update pdb
One should prepare around 100Gb of storage size for the PDB.
Many aspects of the structural analysis performed by StructMAn are based on residue interaction network (RIN) representations of protein structures. Calculating them from a PDB structure is tedious. When using StructMAn a lot or on big datasets it can be advantageous to provide a database of precomputed RINs. Similar to the local instance of the PDB this database has to be stored in structman/resources/. It can be automatically generated using a local instance of the PDB as basis:
structman update rindb
Both databases can also be generated at the same time:
structman update pdb rindb
The calculation of the RINdb takes a lot of computational power, one should provide as many cores as possible. With 48 cores one should still expect one day of runtime. The database requires around 35 Gb of disk space
Protein identifier mapping and protein sequence retrieval utilizes services from UNIPROT. A local instance can be generated. Therefor multiple raw files have to be downloaded (provide temporarily 100Gb of disk space) and a SQL database is constructed out of it. The database requires around 50Gb of disk space. The index table construction needed for fast retrieval operations is a tedious task, thus the database construction takes around one day of computing time.
structman update mapping_db