Where communities thrive


  • Join over 1.5M+ people
  • Join over 100K+ communities
  • Free without limits
  • Create your own community
People
Activity
  • Apr 22 14:09

    Cradoux on master

    update scoring to use volume (compare)

  • Mar 25 15:37

    Cradoux on master

    fixed grid.minimal (compare)

  • Mar 23 19:42

    Cradoux on master

    break before threshold 0 cavity (compare)

  • Mar 18 16:09

    Cradoux on master

    remove g.minimal from extraction (compare)

  • Feb 22 17:02

    Cradoux on master

    added extract all volumes funct… (compare)

  • Feb 01 08:31

    Cradoux on master

    updated requirements (compare)

  • Jan 29 13:51

    Cradoux on master

    better error handling for volum… (compare)

  • Jan 13 10:54

    Cradoux on master

    get crossminer feat db from env… (compare)

  • Jan 12 16:32

    Cradoux on master

    protein scoring uses crossminer… Merge branch 'master' of https:… (compare)

  • Dec 17 2020 11:17

    prcurran on master

    Update tutorial.txt creating t… Update tutorial.txt Updated do… Update grid_extension.py re-or… and 9 more (compare)

  • Dec 17 2020 11:17
    prcurran closed #66
  • Dec 17 2020 10:14
    jasonccole review_requested #66
  • Dec 17 2020 10:13
    jasonccole opened #66
  • Dec 17 2020 10:06

    jasonccole on ccdc_2021_api_updates

    Update tutorial.txt Updated do… Update grid_extension.py re-or… Update calculation.py reorder … and 7 more (compare)

  • Dec 16 2020 12:04
    exsCradoux added as member
  • Dec 16 2020 11:15

    jasonccole on ccdc_2021_api_updates

    Update tutorial.txt creating t… (compare)

  • Dec 16 2020 11:15
    jasonccole added as member
  • Nov 24 2020 09:33

    Cradoux on master

    score mols uses atom volume (compare)

  • Nov 11 2020 14:10

    Cradoux on master

    added grid.copy() to sphere sco… (compare)

  • Nov 05 2020 10:18

    Cradoux on master

    added tmp dir prefix for hotspo… (compare)

Pedro Lacerda
@pslacerda
Hi Pete
Peter Curran
@prcurran
Hello!
Peter Curran
@prcurran
Pedro, I've just added the pymol finish_launching method to the template string module so future results should always have this in the pymol_results_file.py. I've pushed the change.
Thanks for pointing this out :)
Pedro Lacerda
@pslacerda
Hi
I would take some time learn the hotspots project right now, so setup my IDE.

Is the same as yours and seems that some conflicts are happening!

12:08 Load Settings
Cannot load settings from file '/home/pedro/Documents/hotspots/.idea/misc.xml': ParseError at [row,col]:[6,2]
Message: The content of elements must consist of well-formed character data or markup.
Please correct the file content

12:08 Load Settings
Cannot load settings from file '/home/pedro/Documents/hotspots/.idea/fragment_hotspot_maps.iml': ParseError at [row,col]:[5,2]
Message: The content of elements must consist of well-formed character data or markup.
Please correct the file content

do you mind if i remove the .idea directory from git?

ohh, it is already on gitconfig. So I'll just remove, alright?

Probably the .idea directory will suddenly disapears and some setup will be needed

Pedro Lacerda
@pslacerda

did you restored README.md?

I agree that .rst is a bit awkward but is very easier (ie possible) to create nice documentations (eg with references etc) like a documentation book

Pedro Lacerda
@pslacerda
do you mind to reformat the code with black and isort?
i like them very much. Not always do what we want, but free us the need to keep the code formatted ourselves
Pedro Lacerda
@pslacerda
Hi
Pedro Lacerda
@pslacerda
I expect to try Hotspots right after carnival, here in Brazil is a very big holiday with no way to came to the university! Then I'm supposed to try it a bit more.
Peter Curran
@prcurran
Hey @pslacerda , how are you getting on with the Hotspots API?
I would be very keen to hear about how you are using it, if you'd like to arrange a skype my email is pcurran@ccdc.cam.ac.uk

Sorry I didn't reply to the messages above!

.idea completely fine removing that. It's in the gitignore but for some reason made its way into the repo

Yes happy for reformatting code with black and isort, maybe branch that change incase things break :)
Peter Curran
@prcurran
With the README.rst, I reverted the changes as the links no longer worked. Hope that is okay?
Thanks again for your input!
Arian Jamasb
@a-r-j
Hi Peter, great project. I'm having some trouble running this on some structures. Some of them take a very long time to compute - for instance 6vyb has been running for over 15 hours without progress. Admittedly, it is a large structure. Is this to be expected? It runs fine on most other structures I've tried.
Peter Curran
@prcurran

Hi Arian! Thanks for your interest.

For large proteins runtimes can be very long. In these cases it might be worth running on a cavity by cavity basis. Cavities can be detected using the CCDC Cavity module link to docs below:

(https://downloads.ccdc.cam.ac.uk/documentation/API/modules/cavities_api.html?highlight=cavity#module-ccdc.cavity)

The runtime can also be reduced by decreasing the sampling threshold and/or number of rotations generated. You can change these parameters in the Runners's Settings class and parse the Settings object to the from_protein() method on calculation.

e.g.
Runner.Setting.polar_translation_threshold = 14
Runner.Settings.apolar_translation_treshold = 14
Runner.Setting.nrotations = 3000

However, altering the sampler settings in this way will reduce the amount of sampling carried out and could affect the precision of the calculation, so proceed with caution.

If there is output on your console, it implies SuperStar is still running (in which case the sampling step will not yet have been carried out). Therefore, i'd give a cavity run a try and see if that improves things.

Since SuperStar is a product of the CCDC we are limited to what we can change in the scope of this project. However, if SuperStar performance is limiting the work you'd like to carry out please feedback to:
support@ccdc.cam.ac.uk
as this helps to build a case for optimising the SuperStar algorithm.

I hope this helps and do let me know how you get on!

All the best, Pete

Arian Jamasb
@a-r-j
Many thanks, Peter! This is very helpful. Yes, I think SuperStar is the rate limiting step so to speak. The sampling runs quite speedily. Do you have some rough indication of how runtime scales with input size? Runtimes have ranged from 10 minutes to 60h so far which is obviously a huge range.
Cradoux
@Cradoux
Hi Arian, I would definitely suggest running this on a per cavity basis for such a large structure. I have had some structures that extend into >12 hours, at which point I would kill the job and repeat on a per cavity basis
Arian Jamasb
@a-r-j
Hello again, I'm looking to visualise the outputs using MolStar. I'm struggling to figure out a may to convert the .grd files into a readable format. Do you have any suggestions?
Peter Curran
@prcurran
Hi Arian, Sorry for the slow reply. I've just had a play around with MolStar and writing out the FHM results to .ccp4 works fine. :)

settings = HotspotWriter.Settings()
settings.grid_extention = ".ccp4"

with HotspotWriter(outdir, settings) as w:
w.write(hr)

Arian Jamasb
@a-r-j
Hi Peter, thanks for this! It looks like you also have to specify the grid_setting in the HotspotWriter class. Not sure if this is intended functionality
e.g:
with HotspotWrite(outdir, grid_extension='.ccp4', settings=settings) as w:
Peter Curran
@prcurran
Yea that's not intentional. I think we decided to parse the grid_extension var directly (i.e not in the settings class) since it's a common thing people want to change). I'll add this to the list of changes :) Thanks, Pete