model making
February 8, 2010 by Steve · Leave a Comment
A recent paper by Dalal et al: Rapid digital game creation for broadening participation in computing and fostering crucial thinking skills. International Journal of Social and Humanistic Computing, 2009; 1 (2): 123 investigated how the process of creating video games can boost students’ critical and creative thinking skills as well as broaden their participation in computing. The paper indicates that using rapid prototyping with visual representations and event driven behaviors develop creative thinking as well as computing skills.
Recently I have been going through the process of building a computer model of the gene circuitry I am building in the lab. I have found this activity to be quite beneficial, not only in the simulation tool I am creating but the process of creating it. I have found that simply thinking about the interactions of the various components and drawing out visual representations on paper, even without any sophisticated computer software helps to build on the strategy being undertaken in the lab. The approach highlights missing pieces of data and potentially important experiments to characterize the circuits behavior.
For the programming/modeling work (after a long time hammering away at Mathematica and Matlab) I decided to use the Copasi simulation and analysis software package as this gives me the fastest route to prototyping the gene circuit and provides the option to output as sbml which could later be implemented in Matlab if required. Copasi has enabled me to jump straight into the modeling from the point of view of the reactions and at the moment mass action kinetics seems to be sufficient to represent the genetic interactions. Copasi also enables optimization of the model. Signal to noise ratio is key to the circuitry I am building so I am able to optimize the model parameters with the input and output as objective functions. Copasi can then implement a number of deterministic and stochastic optimization algorithms to evolve the system towards the favored output.
The results of this approach have provided some hypothesis about the properties of the gene circuit components in order to achieve the desired output. This has provided a sort of first round iterative design/development cycle and once the lab data has been obtained should provide some interesting validation of the modeling/wet lab approach to the project.
More information on optimization algorithms can be found in Mendes & Kell (1998) Bioinformatics, Vol 14, 869-883.
mathematized
November 17, 2009 by Steve · Leave a Comment
Started a pet project trying to code up the equations from the Mackay and Glass, Oscillation and Chaos in Physiological Control Systems (1977) paper. I need to learn some maths, and some maths programming to broaden my skills beyond wet lab molecular biology. Copasi is ok but it’s reaction based and I worry a point and click interface wont get me a job when I graduate. However, I am rapidly lost in Matlab and Mathematica. I chose Matlab for compatibility and support in the MIB community, but drifted into Mathematica as it seems easier to program using symbolic maths. Both are ok with algebra, then I just get blown away doing anything dynamic like ODE’s. If anybody reads this and has any good tutorials for programming up ODE systems in either/any computing language please let me know :/
Ubuntu 9.10 countdown
October 27, 2009 by Steve · Leave a Comment
Karmic Koala is out, and it’s biting at the heals of Windows 7 and Snow Leopard.
“You are capable of choosing your own destiny. The question is, which path will you choose?”
I’ve been running Windows 7 on a dual boot with Linux Mint for the duration of the Windows 7 beta program. Windows 7 has proved to be a worthy successor to Windows XP, and dragged Microsoft from the pit of despair they hurled themselves into with Windows Vista. In the wake of Vista and the iPOD, the Mac community was able to plough forward with OSX, gaining increasing ground with the macbook pro and air, and more recently had an update from Snow Leopard providing increased speed and some tweaks to their interface. Ubuntu peaked with Jaunty Jackalope and finally released a Linux distribution that gave the Linux community a foothold in the home desktop market. Hardcore Linux users balk at Ubuntu and the heresy of proposing a pre-built linux OS. It has been the long tradition that Unix users compile their own OS and applications, custom built for their own machine, learning the code along the way. There is a steep learning curve for many computer enthusiasts who seek the stability and speed of a Unix OS, but lack a lifetime of dedicated study in the dark art of the command line. Ubuntu broke the mold and provided a generic ready to roll Linux OS that was completely free of any proprietary drivers and codecs, so it could be distributed in any country to any hardware and enable everybody to use a computer regardless of wealth.
Ubuntu 9.10 (Karmic Koala) was released yesterday and continues to build on the improvement of intrepid ibex and jaunty jackalope. I’ve installed a copy of the 32bit desktop edition on a Dell XPS M1530. Installation was much the same as previous version 9 incarnations, with some graphic improvements. Drive partitioning was straight forward enough and it will sit alongside the existing Windows 7 partition without a problem. The OS installed in about 15 minutes and grub picked up the Windows installation to dual boot. I don’t like the new login window. The old version was more streamlined, with the current version requiring me to click on my account and enter my password, even though I’m the only user account on the machine. They also seem to be increasing the number of animated splash screens during the login, where 1 is unnecessary in my opinion. Get to the desktop and stop wasting computer cycles on spinning logos and flashing lines.
Ubuntu comes with no proprietary drivers so I had to install the latest nvidia driver set and download broadcom wireless drivers. This would be a major problem if I didn’t have ethernet internet access as straight away my laptop can’t connect to the internet. Fortunately I do have a wired connection so I could just grab them from the snaptic repository. This wouldn’t be obvious to the casual computer user though and would most likely, from their perspective, brick the machine and send them back to Windows / OSX. After this brief inconvenience I added in the restricted repositories and downloaded all the 3rd party codecs for multimedia playback and I have a ready to go OS again. DVD and MP3 playback is excellent in Totem player and rhythm box. I downloaded Exaile and VLC for myself but the pre-installed media playback tools are excellent. Ubuntu 9.10 comes with Firefox 3.5 which is much faster than 3.x previous versions, with improved Java engine. There is no pre-installed Thunderbird or Sunbird which I don’t understand (considering Firefox is there), instead choosing Evolution, which I haven’t used so can’t comment on. Ubuntu comes pre-loaded with Office 3.1.1 which is more than sufficient for word processing, spreadsheet, and presentations, supporting open document format and mostly compatible with Office 2003 (some compatibility with Office 2007, but personal experience has been bad with powerpoint 2007). Pidin IM has been replaced with empathy in this karmic. Empathy is compatible with all the major IM clients, as well as pidgin’s facebook chat plugin. 9.10 uses the ext4 filesystem as default providing faster data access speeds than previous ext3 and Windows ntfs. Copying files between ntfs and ext4 is noticeably quicker, and it boosts the OS bootup time considerably. 9.10 has also implemented a new “software centre” in place of the add/remove programs in 9.01. This is a streamlined equivalent of the old application, grouping software into easy to navigate sections for first time users. Personally I use the synaptic package library but it’s very useful for new users to find their way around the huge library of additional software that is available.
One of the newer implementations has been Ubuntu one – an online storage drive from canonical providing 2gb of free storage. I found this service to be less reliable than dropbox, which can be installed as a 3rd party application. It had some problems syncing my latex documents in my home folder when I made frequent changes. It was a beta application in Ubuntu 9.04 though and I might give it another go in 9.10. It’s still not bad to have 2gb free synchronized file storage out of the box for free, and an interesting feature that isn’t included in Microsoft or Apple’s OS.
For the complete list of the updates in Ubuntu 9.10, check out the new features page at Ubuntu.com here.
So for the grand price of £0 you get a ready to rock operating system in 32 and 64bit flavors that can (after a little bit of downloading) play high definition movies, music, do all your graphics and office work, surf the net, and send your e-mail. I hope Ubuntu continues to develop in the future, and spin off distributions like Linux Mint build on the Ubuntu base providing custom variations for any perceivable application. Hardcore Linux users might balk at the ethics of Ubuntu but it provides a shallower learning curve for new users like myself to experience the Linux world and break the monopoly of Microsoft and Apple in the home computing market. Once familiar with the environment it enables more serious users to move into the realm of Arch and even Gentoo, eventually compiling their own OS and joining the ranks of the command line battle hardened linux community.
Socialist computing has arrived.
Evolution and Design of Biomolecular Systems A workshop exploring the relationship between systems biology and synthetic biology
October 14, 2009 by Steve · Leave a Comment
We’re on pre-flight checks for Evolution and Design of Biomolecular Systems conference in Mallorca, Spain.
I’m looking forward to meeting other synthetic biology researchers. A few of the conferences I’ve attended so far as part of the systems biology program have been heavy on dry scientists and there has been little interest in wet lab experimentation, leaving me somewhat disjointed from the systems biology community. I am hopeful that the workshops will enable me to gain advice and incite into the work I am doing, as well as the future aims for the field of synthetic biology. I’ll try and write something while I’m there if I can get internet access.
OfficeTab
October 13, 2009 by Steve · Leave a Comment
Useful tool of the week: OfficeTab
OfficeTab adds a tabbed interface to Word, Excel, and Powerpoint. Rather than having separate windows with different files, this little 3rd party Office add-in opens them as tabs within the main application making it much easier to flick between different windows. The authors site is in Chinese but the download links are in English.
September 21, 2009 by Steve · Leave a Comment
Today is 25 years of David Braben and Ian Bell’s Elite video game. This isn’t really systems or synthetic biology, but back in 1984 it was the inspiration of Cambridge students David Braben and Ian Bell that captured the imagination of a generation who spent the next 25 years carving a name for themselves in the Elite universe. Published by AcornSoft in September 1984, Elite has long been considered one of the best video games of all time and inspired the whole “space trading” gaming genre with games such as Eve online, Freelancer, and the X series continuing the tradition. Elite was the first video game to create an “open universe” for players to explore and follow their own path as they chose, creating the concept of “sandbox” gaming popularized by the likes of Grand Theft Auto and Morrowind decades later. Long before the “RPG” scene emerged from dungeons and dragons with class based character development Elite enabled players to choose their own path as explorer, trader, pirate, bounty hunter, mercenary pilot, or all of the above literally doing as they pleased within a dynamic universe. Players could upgrade their ships to haul more cargo across the galaxy or more effectively vaporize the competition on their quest for “Elite” combat status. Explorers could roam over 2000 worlds across 8 galaxies, while budding entrepreneurs could chose to find lucrative trade lanes between planets trading between industrial and agricultural economies slowly making their fortune. Meanwhile more nefarious players could become bandits and pirates, stalking rich systems for cargo laden traders, living as a fugitive from deadly Viper police patrols, seeking refuge in anarchic systems dealing in contraband narcotics and slaves. Whatever your persuasion, Elite had a corner of the galaxy for you.
Many of the aspects of Elite sound routine in 2009, with the likes of Eve online, World of Warcraft, and Starwars Galaxies, but all this was available 25 years ago on an 8bit home computer with 32kb of memory and no internet and was an enormous computing achievement. Elite contained 8 galaxies, each with 256 planets which had to be procedurally generated. A single seed number was run through a fixed algorithm the appropriate number of times and created a sequence of numbers determining each planet’s complete composition (position in the galaxy, prices of commodities, and even name and local details — text strings are chosen numerically from a lookup table and assembled to produce unique descriptions for each planet). This means that no extra memory is needed to store the characteristics of each planet, yet each is unique and has fixed properties. Each galaxy is also procedurally generated from the first. Elite was a technical masterpiece of it’s time and remains unbeaten in it’s achievements and success.
For those who have never experienced the Elite way of life check out the 25th Anniversary website, or for those who fancy dusting off their old combat skills Oolite continues the legacy of Elite with community maintained content and development
… at least while a generation awaits Elite IV.
COPASI – Complex Pathway Simulator
September 9, 2009 by Steve · Leave a Comment
A few weeks ago I attended a modelling and simulation Copasi workshop run at the MIB by Professor Pedro Mendes. I had attempted to blog about it previously but lab work got in the way.
The workshop was a 3 day event detailing all aspects of the Copasi software, much of which can be found in chapter 2 of the “methods in molecular biology” (2009) Volume 500. 1-43, available as a preview here. (There is also a publication associated with Copasi by Hoops et al. (2006) Bioinformatics 22, 3067-74). I’m no mathematician, so my description of Copasi wont be the most accurate! For me, Copasi is a graphical user interface into the world of mathematical modelling providing an immediate step up in to the capabilities of Matlab and Mathematica armed scientists without requiring particularly large amounts of experience of programming or modelling. The software forms a fundamental toolkit of everything a biologist, or mathematician/computer scientist, needs to build models of systems of reactions and run simulations on them. You can enter your reactions using symbolic algebra equations such as those found in many standard biochemistry textbooks, or directly as systems of ODE’s so it is familiar to both wet and dry scientists. A large number of standard enzyme kinetics equations are available when creating your model such as Michaelies Menten types and hill equations as well as all kinds of inhibitor-substrate relationships, and the ability to enter your own.
At its most basic you can input reactions between species of compounds using symbolic algebra and then create plots of the behaviour or those species as they react together in your system over time. You can quickly gain a grasp however of the underlying power of Copasi when you begin to make more sophisticated enquiries of your system. Using the graphical interface you are able to perform a range of systems biology / modelling techniques from finding steady states, to metabolic control and sensitivity analysis. The real power behind Copasi comes from the advanced features however, particularly the parameter scan which is currently not available in any of the equivalent simulation tools and would require substantial programming experience in Matlab or Mathematica. The parameter scan enables you to set a certain parameter at a range of values and repeatedly run the simulation, plotting the output from each iteration. This, for me was a hugely powerful tool as you can test your model under a range of initial conditions, in a high throughput manner.
Copasi is also capable of performing parameter estimation, which allows you to input laboratory data into your model as parameter values and then fit your model parameter values to the data to “reduce the distance” between your model and your observations and reproduce in vivo representative behaviour. In addition, there are a number of optimization algorithms built into Copasi that can optimize your model towards an objective function, or find conditions under which the model behaves in some particular way. There are a wide range of algorithms pre-programmed into Copasi for these tasks including evolutionary programming, genetic algorithms, particle swarms, Praxis, Hooke and Jeeves and more. For the even braver modellers, Copasi can be used in conjunction with Gepasi, an older relative of Copasi that can be used to run multiple simulations simultaneously. For example, you can have multiple copies of a model representing a culture of interacting cells or systems and run multiple simulations on multiple interacting models!
Models can be imported and exported in xml and sbml format and ODE’s can be exported in LaTeX and MathML formats for transfer between different applications. Models from biomodels.net can be imported directly into Copasi and there is a feature to update model details from the Miriam database. There is also a command line version of Copasi that enables high throughput “automated” modelling processes to be run. The Copasi group is also working on a web interface enabling scientists to access the software through a web browser interface.
The comprehensive tool set available in Copasi provides a hugely powerful tool for the budding systems biologists to immerse themselves in the field of mathematical modelling and perform some fairly rigorous and comprehensive modelling techniques without prior experience of complex mathematical programming. It can also produce data for the biologist with a minimum of mathematical knowledge, providing some interesting incites for experimental hypothesis generation.
Copasi is available as open-source, and free for academic research from http://www.copasi.org, and is under continuous development by a core team of programmers as well as a community of users interacting through an active forum. The software is available for Windows, Mac, and Linux. If you’re currently wresting with Matlab or modelling in general, I would recommend Copasi as an excellent starting point to dive into the sometimes intractable world of mathematical modelling, particularly coming from a biological background.
Firefox maintenance – vacuum the sql database
August 24, 2009 by Steve · Leave a Comment
# Open the Error Console: Tools menu/Error Console
# In the Code text box paste this (it’s a single line):
Components.classes["@mozilla.org/browser/nav-history-service;1"]
.getService(Components.interfaces.nsPIPlacesDatabase).DBConnection
.executeSimpleSQL("VACUUM");
# Press Evaluate. All the UI will freeze for a few seconds while databases are VACUUMed
From lifehacker.com
can a scientist be “multidisciplinary”?
During this years inter-DTC systems biology students conference there was a debate session on the subject “can scientists be multidisciplinary?”. The purpose of the session was to debate whether students are better equiped for scientific research by specialising in a particular field or diversifying across fields. I have heard this debate a number of times since joining the Manchester DTC, which I believe stems from an inherent insecurity in the decision to train in the emergent field of systems biology. Many scientists have the desire to label themselves with an identity through their specialism and join other scientists sharing their label in an almost medieval guild like behaviour. These guilds then form alliances between other like minded individuals within and across Universities. This approach has worked well in the existing reductionist scientific world where research is focused on specific areas of interest, enabling funding bodies to quickly identify and recruit suitably skilled scientists to complete their projects. With the advancements in high throughput automation and computer science an increasing number of projects are pushing towards integrative research projects, combining data and experimentation from diverse scientific fields. These advances have culminated recently in the emergent field of systems biology that is driving a new holistic scientific approach through the integration of many scientific disciplines into multidisciplinary research projects. The existing system of training and recruiting specialist scientists has been adapted to encompass a higher level of organisation, utilizing multidisciplinary teams with a central management body coordinating the multidisciplinary communal effort. Problems arising from this new multidisciplinary group working has been in the area of communication with resistance from each of the parties to work with the other, or a resistance to consider the others ways of working, thought processes, or general scientific rationale. The project coordinators have tackled this by continuing the segregation and coordinating from outside each of the groups, minimizing or eliminating the requirement for the component researchers to interact. More recently however, systems biology has promoted a new, more radical approach of training the individual component researchers with the fundamental key skills of their project colleagues, culminating in the development of pioneering University doctoral training centres such as the MIB where classically trained specialists are re-trained in their opposite fields, combining wet and dry lab skills in a single researcher, and the first wave of “systems biologists” are beginning to emerge.
To return to the point of my post, there were representatives on the conference debate panel from Astra Zenica and Pfizer, and the inevitable question “what are you looking for when recruiting a systems biologist?” came up. I found the answer unsurprising, but disappointing. The large pharma companies are searching for mathematicians who can build them models of biological systems that they can use to direct their research. This statement wasn’t followed up by any of the academic researchers. I was left considering the future and relevance of “multidisciplinary scientists” in the wider world. Philosophically, I believe all scientists should be inherently multidisciplinary, enabling them to investigate the world around them and draw conclusion from whichever realm of science that is relevant to the question. I believe that every individual is capable to a reasonable extent to learn anything, and not just “their field” with the only barrier often being attitude and effort. However, in the real world you have to pay the rent and put food on the table, so the starving scientist must match their skill set with the market demand. Current scientific infrastructure in academia and the private sector is set up to employ the person who has the specific skills to complete whatever task is required that their existing assets in the organisation cannot. A multidisciplinary scientist therefore could be considered a “jack of all trades and master of none” and be continuously out-competed by specialists. Existing academic and industry development frameworks are still configured to develop individuals in a specialist role and multidisciplinary individuals don’t fit in their human resources skills matrix, making employment and development impossible to place within their company hierarchy. In addition, the current “group approach” of systems biology means that multidisciplinary thinking is often only required from those coordinating the project, and not the component individuals facilitating the work, and could be interpreted as an unnecessary diversion, deflecting from the development of a key specialism that would be of value to a potential employer. It gave me the feeling that the pharmaceutical industry is funding and promoting systems biology development with the aim of using it as a vector to cherry pick mathematicians that can be re-programmed with sufficient biological knowledge to adapt their mathematical skills to simulation of biological interactions, and the development of wet lab scientists an unfortunate by product. Mathematicians, after sufficient training in systems biology can manage the wet lab scientists and systems biology projects sufficiently to deliver the data required to populate the models and drive the biological hypothesis generation, with wet lab scientists a part of the process that is yet to be automated.
Just to qualify the above, I don’t intend for this post to sound of the “sucks to be you chained in the lab” general bench scientist rant. I believe that systems biology is a transitory scientific field born out of the lack of in situ data to build accurate real world simulations of biological systems. The wet scientists are employed purely to gather this data to populate the mathematicians models, which the mathematicians then use to generate hypothesis and direct the research. It is my belief that mathematicians are the multidisciplinary scientists in systems biology, and must be multidisciplinary for it to continue. The biologists on the other hand must continue to focus on specialisation to develop new measurement techniques, physically obtain the required information, and critically to retain their market value and employment. I believe that biological scientists in systems biology will eventually diversify into either their original fields, or new ones particularly synthetic biology, while systems biology retain the mathematicians and to a certain extent chemists who follow the computer science/simulation road. I believe that adopting a multidisciplinary approach for systems biology is essential for the mathematicians but this approach is detrimental, if not career suicide for biologists, who would be better directed towards classical fields, or for those seeking something new and akin to systems biology – synthetic biology.
I would like to gain comment on this post. I intended for it to be a bit reactionary and maybe inflammatory so I can perhaps incense some response and be proved wrong. I joined the systems biology road trip with the hope of gaining new skills that would drive a new career in biological sciences, and I am hoping for it not to have been a mistake! 
2009 Inter-DTC Systems Biology conference – Manchester MIB
June 27, 2009 by Steve · Leave a Comment
This year’s Inter-DTC systems biology students conference was held at our own Manchester Interdisciplinary Biocentre. Manchester was following on from Warwick’s excellent event last year on their home campus, and had invited a range of speakers from Manchester, Warwick, and Oxford’s systems biology doctoral training centers.
The talks were divided into sections covering many aspects of systems biology ranging across application to theory. There were a number of interesting keynote speakers. Dieter Weichart talked on “from ‘omics’ to systems biology, and David Broomhead gave an intriguing lecture on applying fractal maths to simulate the complex composition of the cell cytoplasm. Professor Broomhead has an excellent way of communicating complex theorem to the most mathematically inept (me!). Pedro Mendes presented a history of simulating biochemical reactions taking us back to the early days of punch cards and drive through size computers crunching away and amusingly producing the same graphs your quad core Intel 8gb DDR-3 powered nvidia SLi carbon footprint behemoth produces 
The students presented a diverse range of projects representative of the style of the doctoral training centre in each University. Oxford University gave an impressive visual presentation of their progress modelling blood vessels within the heart, building on the existing model of the heart by Noble. There were a number of talks ranging across modelling cell signalling, tracking cell movements, and bioengineering of microorganisms for potential biotechnology applications. In addition, there was also 2 interesting talks that diversified from the cell biology / microbiology themes, on modelling / “decoding” epilepsy EEG seizure data to provide new understanding of the underlying cause of the condition and provide potential technologies for predicting and managing seizures. One talk focused on absent seizures while the other focused on grand mal seizures, bother using different mathematical and statistical methods to search for patterns in the inherantly complex and chaotic brain activity.
The event was a great way to encourage students to communicate between systems biology DTC’s as well as building links for current and future research. I enjoyed the opportunity to meet with students working on areas similar to my own, as well as their supervisors who could give their own incite on the work that I was doing. This helps to broaden your understanding, as well as gain fresh ideas from people outside of your usual research group. After the conference we went for an excellent meal hosted by Professor Westerhoff and had a chance to interact in a more relaxed social environment.
Next year’s conference hopefully will be hosted by Oxford University, which would be a unique and priviledged opportunity to visit their prestigious campus, and experience their own blend of systems biology research.