This forum is aimed towards anybody who is working with the the MembraneEditor. Installation instructions are found here. Any questions, remarks or critics are welcome. Also this is a chance to get in contact with other users.
Post Reply
User avatar
bjoern
Site Admin
Posts: 211
Joined: 20.03.2006, 15:01
Contact:

Bug Report MembraneEditor

Post by bjoern » 19.01.2013, 11:02

If you find any bugs while using the MembraneEditor, please add them here, and we will try to fix it in the next release.

Philipp
Posts: 7
Joined: 30.07.2014, 09:57

Re: Bug Report MembraneEditor

Post by Philipp » 30.07.2014, 10:15

Hi,

I am playing around with the Cm2 MembraneEditor and got stuck while trying to model micells and small liposomes.

Everytime I try to create a micelle using the Vesicle Builder I get an error message (java.lang.NullPointer Exception). Lipids were assigned to EC. By assigning lipids to IC they are added to the underlying sphere, but the process pends in the initial adding lipids phase and I can not stop the process.

Is it normal that I can not interrupt the inital adding lipid process? In my experience is this only possible after the inital process is finished and the lipids are being shaken but it would be nice not to have to kill the whole programm again and again.

Best,
Philipp

User avatar
bjoern
Site Admin
Posts: 211
Joined: 20.03.2006, 15:01
Contact:

Re: Bug Report MembraneEditor

Post by bjoern » 30.07.2014, 12:27

Hi Philipp!

Yes, you are right - the former version of the VesicleBuilder did not work for micelles.

I just uploaded the new version of the VesicleBuilder which supports also to pack only the external layer.

https://www.cellmicrocosmos.org/Cmforum ... f=35&t=676

Just activate fill only EC.

By the way, make sure that if yo have multiple versions of the The Vesicle Builder on your hard disk, that the used jar file needs to have the name "VesicleBuilder.jar" - otherwise the MembraneEditor will not find the correct entry class in the jar file.

Let me know if you need more options in the VesicleBuilder or if something does not work.

Best wishes,
Björn

Philipp
Posts: 7
Joined: 30.07.2014, 09:57

Re: Bug Report MembraneEditor

Post by Philipp » 10.09.2014, 13:05

It's me again :D

there are some issues I came across.
  • While editing the conformation of a lipid it happens that atom bonds are being stretched. Sometimes several bonds within the selection that is rotated strech, sometimes it is just the bond that is attached to the center of rotation.
  • You can not set the center of rotation in a selection and you can not deselect the center if you accidentally enclosed it in the selection that is to be rotated.
  • The Wanderer Rookie algorithm is not working for domains.
  • Working with multiple layers and also mutiple domains it seems to be not possible to assign PSCe to all domains. I can assign PSCe to multiple domains on the default layer (I tested up to nine domains), but as soon as I switch to the second layer I am able to assign PSCe to just one additional domain. But this is true only if I skip the default aerea and start assigning at the domains. If I assign PSCe to the default area in the first place and afterwards to one of the domains I am able to assign PSCe to both, the default area and the domain. This also holds true for a third layer. You could argue that a PSCe domain in a PSCe layer makes no sense, but even if you assigne another lipid to the default area on the second layer you are only able to assign one PSCe domain.
  • The Vesicle builder won't start until you assigned lipids to every domain in your setup. Why is there a "cross-talk" between vesicle and planar membrane?
  • "Clear lipids" does not work properly unless you clear everything. Sometimes a built vesicle is not affected, sometimes some lipids remain, while having an additional planar membrane.
  • Are multiple layers and domains an issue for the program? Over time there are misbehaviours like the content of the local database section is also shown in the statistics area or there are some white areas appearing (the RAM should not be the problem. The program (webstart) is allowed to take up to 10 GB of memory and actually it uses 1.3 GB). Those misbehaviours also occure if I try to add a second layer.
Best,
Philipp

\edit
OK, this lipd-domain-assigning-thing seems to be more complex. I just tried to set up two layers each with two domains. I assigned PSCe to both domains in the default layer and tried to assign PEWe to both domains on the second layer and surprise ... I was only able to assign PEWe to one of them. So probably it is not a "you-used-that-lipid-too-often"-issue :/

\\edit
I am testing some issues from above:
  • In the question thread I told about this sandwich-like setup for generatin lipid rings for reconstructing a fusion event in Blender. So it is not working. I tried to set up two layers that are 66 A far apart of each other. Then I created two domains on each layer covering the upper and lower third leaving a space of 50 A. Afterwards I added a third layer in-between the other two and tried to create a vesicle that would be blocked in both Z-directions because of the domains. But I could not start the algorithm because it expects lipids in the other default areas (it says area, but I think this holds true for both layers).
    As a second approach I marked just two domains in the middle of the layers and assigned the lipid to the default areas leaving the domain empty. If then try to built a vesicle in the mid-position layer I get the error message "no lipids could be added available!" and the graphical defects appear until you resize the program.
    To simplify the whole thing I then changed the size of the default layer to be 70 A in Z-direction. Doing so I can get the Veicle Builder starting to build the ring but as soon as finishes the initial adding and starts to add random rotains and so on I can not stop the algorithm. Activating the "gear" under the viewport you can see what is going on at the moment (adding, twisting ...) and this sequence is repeating itself so fast that it looks like doing nothing at all but got stuck in an endless looping. Acutally the lipids in the ring are not moving at all. Maybe this is a fact of the area that is outside the bounding box? I remember that my first vesicles got tapered pols until I figured out that they were bigger than the bounding box. Nevertheless the were build without any problems. I also tried to built just one half of a vescile using this box-limitation-method and got the same endless looping.
  • Graphical defects are appearing almost all the time by now no matter how complex the scene is. I suspect that my Java is the troublemaker. Yesterday I updated Java and thereby all older versions were removed :( But with Java 7 the MembraneEditor does not start so I installed Java 6 with the latest update (u45). I thinks this was a mistake. Never touch a running system :evil:
  • "The Wanderer Rookie is not working on domains" as I stated before is not true. It just can't handle empty domains.

Post Reply