Re: [cellml-discussion] Using CellML to represent huge CellML models:Has anyone worked on this already?
I am working on developing a CellML model (using external code) of transcriptional control in yeast which is 23 MB in size. I hope to eventually do a similar thing for organisms which have much more complicated sets of interactions, in which case this size may grow substantially. so you have 23MB of XML? Cool! Even combining all my models I have less than 7MB, and even then I'm sure that figure includes some simulation results. I guess an interesting test would be uploading it to the model repository to see how that handles such a large model (presuming you have a CellML 1.0 model). If anyone on this list is interested in similar problems (I presume similar issues come up in a range of systems biology problems, whether you are working with CellML or SBML), I would welcome your feedback and suggestions, and perhaps we could collaborate . I really have no idea what an transcriptional control in yeast model looks like, but my initial thought would be to abstract out any similar math and import common declarations - I'm guessing you have already done this if its possible. This creates some unique issues for CellML processing tools: 1) Just parsing the CellML model (especially with a DOM-type parser which stores all the nodes into a tree, but probably with any type of parser) is very slow. it might be interesting to look at doing some simple task to check the performance of DOM vs SAX based tools? I have found in the past that with 500MB fieldML files that the SAX parser used in CMGUI was quite fast at parsing the file - especially if you go from a gzip compressed file. 2) The CellML model might not all fit in memory at the same time, especially if the model gets to be multi-gigabyte. It might be possible to make use of swap to deal with this, but if the algorithms don't have explicit control over when things are swapped in and out, it will be hard to work with such a model. I think if you have a model getting that large then there needs to be some serious thinking about how to handle such models...but generally can't you just let the OS worry about swapping in and out as required? Or would you expect a customised scheme for a particular application to be more efficient? C) Another leaner API, read-only CellML API (perhaps based off the same IDLs, but with certain functionality, like the ability to modify the model, or set mutation event listeners, unavailable). We could add a SAX-style event dispatcher instead, to allow users to save any information they do want from extension elements, which will also not be kept in the model. Comments, white-space, and so on would all be stripped unlike in the current CellML API implementation. Tools which are currently using the full CellML API but only require read-only access (e.g. the CCGS) might be able to just 'flick the switch' and benefit from the leaner API. This would probably be beneficial even for those of us without such large models - especially if it is as easy as flicking a switch to swap between the complete and restricted implementations. Andre. ___ cellml-discussion mailing list cellml-discussion@cellml.org http://www.cellml.org/mailman/listinfo/cellml-discussion
Re: [cellml-discussion] Using CellML to represent huge CellML models:Has anyone worked on this already?
David Nickerson wrote: I am working on developing a CellML model (using external code) of transcriptional control in yeast which is 23 MB in size. I hope to eventually do a similar thing for organisms which have much more complicated sets of interactions, in which case this size may grow substantially. so you have 23MB of XML? Cool! Even combining all my models I have less than 7MB, and even then I'm sure that figure includes some simulation results. My model is entirely generated from experimental data, none of it is written by hand (aside from a one-page script used to generate CellML from the relational database). I guess an interesting test would be uploading it to the model repository to see how that handles such a large model (presuming you have a CellML 1.0 model). It is currently a CellML 1.0 model. I'm not sure I want to break the live Plone, however. I'm not sure it is much use to anyone else at this stage, however. If anyone on this list is interested in similar problems (I presume similar issues come up in a range of systems biology problems, whether you are working with CellML or SBML), I would welcome your feedback and suggestions, and perhaps we could collaborate . I really have no idea what an transcriptional control in yeast model looks like, but my initial thought would be to abstract out any similar math and import common declarations - I'm guessing you have already done this if its possible. My model only has machine-learning external-code in it, it doesn't have any equations at the moment. Just to give you an idea of what it looks like... model xmlns=http://www.cellml.org/cellml/1.0#; name=interactions component name=PAU8 variable name=sig_PAU8 initial_value=0 units=signal_level public_interface=out/ variable name=sig_SUT1 units=signal_level public_interface=in/ variable name=sig_STE12 units=signal_level public_interface=in/ variable name=sig_ADR1 units=signal_level public_interface=in/ variable name=sig_YAP5 units=signal_level public_interface=in/ variable name=sig_RME1 units=signal_level public_interface=in/ variable name=sig_TEC1 units=signal_level public_interface=in/ variable name=sig_SWI5 units=signal_level public_interface=in/ variable name=sig_ARR1 units=signal_level public_interface=in/ variable name=sig_MET31 units=signal_level public_interface=in/ variable name=sig_RLM1 units=signal_level public_interface=in/ variable name=sig_INO4 units=signal_level public_interface=in/ variable name=sig_RAP1 units=signal_level public_interface=in/ variable name=sig_MOT3 units=signal_level public_interface=in/ math xmlns=http://www.w3.org/1998/Math/MathML; applyeq/ cisig_PAU8/ci apply csymbol definitionURL=http://www.bioeng.auckland.ac.nz/people/miller/black_box/k-nearest-neighbours;blackbox/csymbol cisig_SUT1/ci cisig_STE12/ci cisig_ADR1/ci cisig_YAP5/ci cisig_RME1/ci cisig_TEC1/ci cisig_SWI5/ci cisig_ARR1/ci cisig_MET31/ci cisig_RLM1/ci cisig_INO4/ci cisig_RAP1/ci cisig_MOT3/ci /apply /apply /math /component component name=YAL067W_A variable name=sig_YAL067W_A initial_value=0 units=signal_level public_interface=out/ variable name=sig_SPT23 units=signal_level public_interface=in/ variable name=sig_STE12 units=signal_level public_interface=in/ variable name=sig_DAL80 units=signal_level public_interface=in/ variable name=sig_YAP5 units=signal_level public_interface=in/ variable name=sig_BAS1 units=signal_level public_interface=in/ variable name=sig_DIG1 units=signal_level public_interface=in/ variable name=sig_PHO2 units=signal_level public_interface=in/ variable name=sig_HAP2 units=signal_level public_interface=in/ variable name=sig_PHD1 units=signal_level public_interface=in/ variable name=sig_GLN3 units=signal_level public_interface=in/ math xmlns=http://www.w3.org/1998/Math/MathML; applyeq/ cisig_YAL067W_A/ci apply csymbol definitionURL=http://www.bioeng.auckland.ac.nz/people/miller/black_box/k-nearest-neighbours;blackbox/csymbol cisig_SPT23/ci cisig_STE12/ci cisig_DAL80/ci cisig_YAP5/ci cisig_BAS1/ci cisig_DIG1/ci cisig_PHO2/ci cisig_HAP2/ci cisig_PHD1/ci cisig_GLN3/ci /apply /apply /math /component ... Note that the initial_value=0 is a place-holder, I could abstract out my blackbox function calls based on the number of parameters (it is variable, from 1 through to 41, in this case, although there is no theoretical limit on how many putative transcription factors could affect a signal). However, I suspect that this would not solve the performance problems (it takes
Re: [cellml-discussion] Dimensional consistency andunitsconversions (was [Fwd: Re: ten Tusscher model])
So one way out to avoid having to hope that software does the right thing is to make it compulsory that there is units consistency for each dimension across all variables (defined in variable elements) in a CellML component so that the only units conversions that need to take place are at the interfaces. Quantities that are dimensionless after simplification aren't going to affect dimensional analysis of the math, so we would be safe there. On 4/23/07, David Nickerson [EMAIL PROTECTED] wrote: So where does the problem lie? This says that all you supposedly dimensionless constants should have units. Does it need to be clearer that you are not allowed to simplify them out into dimensionless yourself? yes - I think this is the issue. Also that tools shouldn't simplify them into dimensionless before doing the multiplication and/or units consistency checking. Also. What does it matter that some software simplifies them out before multiplicating them? So long as it checks units consistency prior to simplifying them (if it really needs to do that anyway) then the result should be the same. yep - that is the key. The units must be there when an application checks units consistency. Andre. ___ cellml-discussion mailing list cellml-discussion@cellml.org http://www.cellml.org/mailman/listinfo/cellml-discussion ___ cellml-discussion mailing list cellml-discussion@cellml.org http://www.cellml.org/mailman/listinfo/cellml-discussion
[cellml-discussion] Proposal: Refactoring the CCGS into smaller, re-usable components
Hi, I have been wanting to use some functionality present in the CellML Code Generation Service, for a quite different type of problem (not a system of ODEs). The problem is, the functionality isn't exposed. I think there is actually quite a lot of useful functionality in the CCGS which could be exposed. I am therefore proposing that the CCGS be split into a number of smaller, independently useful components... 1) CUSES: The CellML Unit Simplification And Expansion Service. This service will allow a string, describing a unit, together with a component or units element in which the unit appears, to be passed in. It will then return another object, which represents the unit in a special canonical. This canonical form will consist of an ordered sequence of base units (for some unique ordering of base units). Base units will include built-in base units as well as user-defined base units. The canonical sequence will also carry multipliers, offsets, and exponents for each base unit. It will be possible to compare to canonical forms of units. There will also be an option for whether to combine exponents when units have a different multiplier, so that, for example, millimetre . (millisecond ^ -1) . microsecond will only be simplified to micrometre if you turn the option on. I think it is this specific type of simplification that has lead to the recent debate over units. It is hoped that this module will be useful for editing tools, validators, and code generators, as well as any other software that needs to worked with units. If you work on such software, let me know if this sounds useful. 2) CeVAS: The CellML Variable Association Service. This service provides an efficient way to find all the variables which are connected to each other, even if they are different components, and perhaps in different imported models. It also allows annotations to be made per actual variable (i.e. group of connected variables), and it uses CUSES to compute conversion factors and offsets needed to convert between one CellML variable and another one connected to it. This should be useful for a range of simulators and code generators, hopefully even ones which are not ODE-based. 3) MaLaES: The MathML to Language Expression Service. This provides facilities for translating individual MathML expressions into code in a specific language. The code will provide a programmatic interface for setting up tables used to drive the code generation, which will aim to support a range of common languages (earlier FORTRANs being the hardest, due to the line-length restrictions). The code will use CeVAS annotations to look up the name of the variables, so you will be able to use arbitrary variable names. This service will also be able to determine certain information about the mathematics, such as returning a set of variables on each side of the equation, and determining whether external code is used. The CCGS will keep the code used to determine the order of the expressions, as well as generation of the actual expressions. I welcome any opinions on whether this is useful, suggested improvements, and so on. Best regards, Andrew ___ cellml-discussion mailing list cellml-discussion@cellml.org http://www.cellml.org/mailman/listinfo/cellml-discussion
[cellml-discussion] some repository stats
Hi Folks, I have now checked every model in the repository in PCEnv and recorded whether they run, and if so, whether they integrate, or if they don't run, what the error message is. The great majority contain undefined variables. As of 24/04/07: Out of a total of 250 different models (I only counted once, so make that plus or minus a few,) 58 models are functional in PCEnv and have therefore made it to a one star rating. That's 23.2 %, which isn't too bad. There are 412 files in the repository, so that equates to an average of 1.65 versions/variants per model. Most of the variants got uploaded automatically from the old repository as versions, so that's something I'll fix eventually. As of now, a star next to the model name in the repository list definitively means that at least one version of that model runs and integrates in PCEnv. James ___ cellml-discussion mailing list cellml-discussion@cellml.org http://www.cellml.org/mailman/listinfo/cellml-discussion