rsilvabioinfo
commented
4 years ago Hello DF,
thanks for your contact.
I imagine that when you say you submitted a database you mean you tried the formatting tool: http://dorresteinappshub.ucsd.edu:5002/upload
The code for both, in house formatting https://github.com/DorresteinLaboratory/NAP_ProteoSAFe/blob/master/formatdb/formadb.ipynb
and for the webserver https://github.com/computational-chemical-biology/formatdb
are available. If you get to the code, you will notice that we only retrieve class assignments for structures already classified, and it is still quite slow (classifier server also limits the number of queries now).
If you need help, you can directly email me the structures and I will do the queries for you.
Cheers, Ricardo
Em qua., 20 de nov. de 2019 às 16:09, dlforrister notifications@github.com escreveu:
Good afternoon. Not exactly an issue but a question.
Our group has generated several large theoretical libraries based on combinatorial chemistry for the common structures/substitutions we anticipate in our dataset. We'd like to use these in NAP workflow, but when uploading .txt files containing smiles and id we get a proxy error. I've subset the dataset to be smaller did not receive the same error.
I am wondering if it's possible to upload these large libraries (90,000 smiles structures~22mb text file)? Similarly, we'd like to upload the Universal Natural Products database (170,000 structures ~13mb total). UNPD, as a large natural product library would likely to useful to other groups as well.
I submitted these large files Yesterday (Nov 19, 2019), via the web portal as a test. I do not know if they are running or bogging down your servers. I can patiently wait for them to run, but if they are bogging down your computers or If this takes up too much computing power on your end. Is it possible to share some of the code that converts the .txt database into the fully classified database that NAP uses? We could likely generate our own formatted databases on our own computing resources.
I just wanted to be proactive and ask how you suggest handling these larger spectral databases.
Thanks -DF
— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub, or unsubscribe.
dlforrister
Good afternoon. Not exactly an issue but a question.
Our group has generated several large theoretical libraries based on combinatorial chemistry for the common structures/substitutions we anticipate in our dataset. We'd like to use these in NAP workflow, but when uploading .txt files containing smiles and id we get a proxy error. I've subset the dataset to be smaller did not receive the same error.
I am wondering if it's possible to upload these large libraries (90,000 smiles structures~22mb text file)? Similarly, we'd like to upload the Universal Natural Products database (170,000 structures ~13mb total). UNPD, as a large natural product library would likely to useful to other groups as well.
I submitted these large files Yesterday (Nov 19, 2019), via the web portal as a test. I do not know if they are running or bogging down your servers. I can patiently wait for them to run, but if they are bogging down your computers or If this takes up too much computing power on your end. Is it possible to share some of the code that converts the .txt database into the fully classified database that NAP uses? We could likely generate our own formatted databases on our own computing resources.
I just wanted to be proactive and ask how you suggest handling these larger spectral databases.
Thanks -DF