I'm downloading PDBe PISA data as advised on the download page, via successive calls with lists of PDB codes such as http://www.ebi.ac.uk/pdbe/pisa/cgi-bin/interfaces.pisa?1kq1,4foj,3ga7,4err etc.
The download time seems to be varying (in the last few sets of 50 interfaces it's been between 3 and 5-6 minutes per set of IDs), so for the 2000+ queries it could take between 100 and 200 hours in total!
I'm wondering if it's common practice to parallelise this in some way (e.g. doing a download on one half somewhere else) or if anybody out there may have a full set that they could share?
I would have considered contacting PDBe but their site makes it pretty clear this is the recommended way to download the information.
P.S. the Python script I wrote for the task here allows for interrupted/resumed download if anyone reading this is seeking to do the same.
1 answer
I've got a more modernized and Pandas-integrated way to do this that I post about here. Snakemake allows you to just provide at a list of PDB identifiers for the script to process. The result will then be Pandas dataframes that can be integrated easily with downstream analysis.
Log in to answer this question.