Re: [Jmol-developers] JSmol as network viewer
On 10/06/2014 08:38 PM, Robert Hanson wrote: > Probably not. I didn't realize we were talking about quite so many nodes. > Still, the UFF field is interesting because it might work with any geometry > -- leaving angles out all together. We might have to add something to get > it to ignore certain energy components. Angles and dihedrals, for instance. > The more I think about it the more I doubt it really makes sense for me to dig deep into UFF and spend a lot of time just to find out that it doesn't really fit to the task. I took a quick look at a paper on UFF from Rappe et al (J. Am. Chem. Soc. 114:10024-10035, 1992) and can't really imagine how I would be able to fit this into my task. Bob, maybe if I would describe some possibly suitable (initial) layout strategies that came into my mind you could give me a hint if it would really be worth the effort to try to make UFF work for networks. For a specific ageing factor there will be for example two types of networks: 1) All observations in which the ageing factor participates and all other ageing factors that are also involved in these observations. This means there will be one central ageing factor node connected to all observations. And each observation will be connected to currently up to six other ageing factors. 2) An iterative combination of type 1 networks for each ageing factor from type 1 into a huge combined network. This will be much more complex with a lot of smaller and larger centers, connected by one or more observations. Finding a suitable layout for type 1 might not be too complicated. I thought about two strategies: 1a) The observation nodes are evenly distributed within a sphere around the central ageing factor node. The other ageing factor nodes are then again distributed evenly within a sphere around their observation node. Nodes with more connections are preferably positioned in the outer parts. 1b) The other ageing factor nodes are evenly distributed within a sphere around the central ageing factor node. The observation nodes are then evenly distributed in the non-overlapping part of a second larger sphere around the central ageing factor. This will be more complicated to achieve than 1a but it has the advantage that it will be possible to group the ageing factors and separately also the observations by user-defined criteria. For example on one side of the sphere could be placed all observations that decrease the lifespan. On the other side could be the observations that increase the lifespan. And in the middle could be the neutral ones. Building type 2 will presumably be more complicated. Only strategy 1a would fit in here I would guess. I don't think that strategy 1b is compatible with type 2, at least not without containing nodes redundantly. (But maybe without a real center...?) My final goal would be to develop a general purpose JSmol-based network viewer (not just AgeFactDB-related). I guess it will be somehow similar to the Jena3D viewer for PDB structures but ideally without a server backend (except maybe as additional information source). Finally I have a related question: Q: Does JSmol allow that an atom is part of more than one residue or would this interfere with something? I ask this because it would enable to use residue-based functions for handling observations. (A residue would consist of the observation node and all connected ageing factor nodes. Ageing factor nodes might then be part of a larger number of residues.) I really hope that this is possible because it would speed up things a lot and also spare me a lot of development time. Regards, Rolf -- Rolf Huehne Postdoc Leibniz Institute for Age Research - Fritz Lipmann Institute (FLI) Beutenbergstrasse 11 07745 Jena, Germany Phone: +49 3641 65 6205 Fax: +49 3641 65 6210 E-Mail: rhue...@fli-leibniz.de Website: http://www.fli-leibniz.de Scientific Director: Prof. Dr. K. Lenhard Rudolph Head of Administration: Dr. Daniele Barthel Chairman of Board of Trustees: Dennys Klein VAT No: DE 153 925 464 Register of Associations: No. 230296, Amtsgericht Jena Tax Number: 162/141/08228 -- Meet PCI DSS 3.0 Compliance Requirements with EventLog Analyzer Achieve PCI DSS 3.0 Compliant Status with Out-of-the-box PCI DSS Reports Are you Audit-Ready for PCI DSS 3.0 Compliance? Download White paper Comply to PCI DSS 3.0 Requirement 10 and 11.5 with EventLog Analyzer http://pubads.g.doubleclick.net/gampad/clk?id=154622311&iu=/4140/ostg.clktrk ___ Jmol-developers mailing list Jmol-developers@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/jmol-developers
Re: [Jmol-developers] JSmol as network viewer
Hi Rolf This is interesting but not what I can really discuss about, but I'd contribute one or two hints: One really nice thing (I've seen it in other tree viewers) would be the ability for the user to pull a node and have the whole thing rearrange dynamically. That's a force field and Jmol already does this with dragMinimze. However, probably not easy to achieve seeing the complexity of your trees (i.e. many "bonds" tied to an "atom"). > Q: Does JSmol allow that an atom is part of more than one residue or > would this interfere with something? Quick answer: no atom can belong to 2 residues. But I would forget about "residues". I'd just define Jmol variables as atom sets and assign each node to as many of them as you wish (using a script). It is as easy to select by user variables / atomsets as by residue. Another possibility is to assign custom atom properties, instead of atom sets. -- Meet PCI DSS 3.0 Compliance Requirements with EventLog Analyzer Achieve PCI DSS 3.0 Compliant Status with Out-of-the-box PCI DSS Reports Are you Audit-Ready for PCI DSS 3.0 Compliance? Download White paper Comply to PCI DSS 3.0 Requirement 10 and 11.5 with EventLog Analyzer http://pubads.g.doubleclick.net/gampad/clk?id=154622311&iu=/4140/ostg.clktrk ___ Jmol-developers mailing list Jmol-developers@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/jmol-developers
Re: [Jmol-developers] JSmol as network viewer
On 10/07/2014 04:50 PM, Angel Herráez wrote: > Hi Rolf > > This is interesting but not what I can really discuss about, but I'd > contribute one or two hints: > > One really nice thing (I've seen it in other tree viewers) would be > the ability for the user to pull a node and have the whole thing > rearrange dynamically. That's a force field and Jmol already does > this with dragMinimze. > However, probably not easy to achieve seeing the complexity of your > trees (i.e. many "bonds" tied to an "atom"). > Yes, Angel. It would be one of the next steps to make the layout directly editable by the user. It could be completely manually or manually-directed like you describe it. But for this you need a reasonable initial layout. I will keep this in mind if I ever get past the first step. >> Q: Does JSmol allow that an atom is part of more than one residue or >> would this interfere with something? > > Quick answer: no atom can belong to 2 residues. > > But I would forget about "residues". I'd just define Jmol variables > as atom sets and assign each node to as many of them as you wish > (using a script). > It is as easy to select by user variables / atomsets as by residue. > Another possibility is to assign custom atom properties, instead of > atom sets. > That is what I do within the 'Jena3D Viewer' for the different mappings (SAPs/SNPs, PROSITE motifs etc.). And from there I know that commands can be very slowly this way. Handling for example several hundred SNPs by using sets already takes quite some time (see for example http://jenalib.fli-leibniz.de/cgi-bin/3d_mapping.pl?CODE=1A00&APPLET=HTML5). And in the network viewer the number of sets would potentially be much higher. It will also require to rebuild a lot of built-in functions to get similar (slower) functionalities. Regards, Rolf -- Rolf Huehne Postdoc Leibniz Institute for Age Research - Fritz Lipmann Institute (FLI) Beutenbergstrasse 11 07745 Jena, Germany Phone: +49 3641 65 6205 Fax: +49 3641 65 6210 E-Mail: rhue...@fli-leibniz.de Website: http://www.fli-leibniz.de Scientific Director: Prof. Dr. K. Lenhard Rudolph Head of Administration: Dr. Daniele Barthel Chairman of Board of Trustees: Dennys Klein VAT No: DE 153 925 464 Register of Associations: No. 230296, Amtsgericht Jena Tax Number: 162/141/08228 -- Meet PCI DSS 3.0 Compliance Requirements with EventLog Analyzer Achieve PCI DSS 3.0 Compliant Status with Out-of-the-box PCI DSS Reports Are you Audit-Ready for PCI DSS 3.0 Compliance? Download White paper Comply to PCI DSS 3.0 Requirement 10 and 11.5 with EventLog Analyzer http://pubads.g.doubleclick.net/gampad/clk?id=154622311&iu=/4140/ostg.clktrk ___ Jmol-developers mailing list Jmol-developers@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/jmol-developers
Re: [Jmol-developers] JSmol as network viewer
On Tue, Oct 7, 2014 at 9:22 AM, Rolf Huehne wrote: > On 10/06/2014 08:38 PM, Robert Hanson wrote: > > Probably not. I didn't realize we were talking about quite so many nodes. > > Still, the UFF field is interesting because it might work with any > geometry > > -- leaving angles out all together. We might have to add something to get > > it to ignore certain energy components. Angles and dihedrals, for > instance. > > > The more I think about it the more I doubt it really makes sense for me > to dig deep into UFF and spend a lot of time just to find out that it > doesn't really fit to the task. I took a quick look at a paper on UFF > from Rappe et al (J. Am. Chem. Soc. 114:10024-10035, 1992) and can't > really imagine how I would be able to fit this into my task. > > Bob, maybe if I would describe some possibly suitable (initial) layout > strategies that came into my mind you could give me a hint if it would > really be worth the effort to try to make UFF work for networks. > > For a specific ageing factor there will be for example two types of > networks: > 1) All observations in which the ageing factor participates and all > other ageing factors that are also involved in these observations. > This means there will be one central ageing factor node connected to all > observations. And each observation will be connected to currently up to > six other ageing factors. > > 2) An iterative combination of type 1 networks for each ageing factor > from type 1 into a huge combined network. This will be much more complex > with a lot of smaller and larger centers, connected by one or more > observations. > > Finding a suitable layout for type 1 might not be too complicated. I > thought about two strategies: > 1a) The observation nodes are evenly distributed within a sphere around > the central ageing factor node. The other ageing factor nodes are then > again distributed evenly within a sphere around their observation node. > Nodes with more connections are preferably positioned in the outer parts. > > 1b) The other ageing factor nodes are evenly distributed within a sphere > around the central ageing factor node. The observation nodes are then > evenly distributed in the non-overlapping part of a second larger sphere > around the central ageing factor. > This will be more complicated to achieve than 1a but it has the > advantage that it will be possible to group the ageing factors and > separately also the observations by user-defined criteria. For example > on one side of the sphere could be placed all observations that decrease > the lifespan. On the other side could be the observations that increase > the lifespan. And in the middle could be the neutral ones. > > Building type 2 will presumably be more complicated. Only strategy 1a > would fit in here I would guess. I don't think that strategy 1b is > compatible with type 2, at least not without containing nodes > redundantly. (But maybe without a real center...?) > > My final goal would be to develop a general purpose JSmol-based network > viewer (not just AgeFactDB-related). I guess it will be somehow similar > to the Jena3D viewer for PDB structures but ideally without a server > backend (except maybe as additional information source). > > The only significance of the UFF would be if the connections you are making in your net have weight. Some "bonds" stronger than others, for example. The UFF could be used if I added a bit to allow turning off various energy components Probably all you would need would be VDW and bond energies. No angles, no dihedrals. > Finally I have a related question: > > Q: Does JSmol allow that an atom is part of more than one residue or > would this interfere with something? > Certainly an atom can only be in one residue. Each atom has a "group" property that points to its residue. > > I ask this because it would enable to use residue-based functions for > handling observations. (A residue would consist of the observation node > and all connected ageing factor nodes. Ageing factor nodes might then be > part of a larger number of residues.) I really hope that this is > possible because it would speed up things a lot and also spare me a lot > of development time. > > Regards, > Rolf > -- > > Rolf Huehne > Postdoc > > Leibniz Institute for Age Research - Fritz Lipmann Institute (FLI) > Beutenbergstrasse 11 > 07745 Jena, Germany > > Phone: +49 3641 65 6205 > Fax: +49 3641 65 6210 > E-Mail: rhue...@fli-leibniz.de > Website: http://www.fli-leibniz.de > >Scientific Director: Prof. Dr. K. Lenhard Rudolph > Head of Administration: Dr. Daniele Barthel > Chairman of Board of Trustees: Dennys Klein > > VAT No: DE 153 925 464 > Register of Associations: No. 230296, Amtsgericht Jena > Tax Number: 162/141/08228 > > > > -- > Meet PCI DSS 3.0 Compliance Requirements with EventLog Analyzer > Achieve PCI DSS 3.0 Compliant Status with Out-of-t
