Naaman Kam et al. (2003) International Worm Meeting "Computer Modeling, Simulation and Analysis of C. elegans Development"

David Harel, Naaman Kam, Michael J Stern, E Jane Albert Hubbard, Amir Pnueli

[International Worm Meeting, 2003]

Na'aman Kam et al. (2004) East Coast Worm Meeting "COMPUTER MODELING, SIMULATION AND ANALYSIS OF C. elegans VULVAL INDUCTION"

Jasmin Fisher, Amir Pnueli, David Harel, E Jane Albert Hubbard, Na'aman Kam, Michael J Stern

[East Coast Worm Meeting, 2004]

We have been developing a new approach to modeling biological phenomena, focusing on the signaling events that influence vulval fate specification in C. elegans . Our motivation is the striking similarity between the methods and logic of developmental genetics with the languages and tools used in the design of complex computerized systems. Vulval precursor cell (VPC) fate specification is an excellent system to test our approach for two reasons: (1) the mechanisms underlying this process are sufficiently understood to allow meaningful models to be created; and (2) the multiple known inputs that determine fate output create a complexity that can adequately challenge the efficacy of the modeling approach. Developmental genetic data are difficult to make tractable for simulation and analysis by computers ("formalize"), since they are often in a condition-result form in which the precise mechanisms linking the conditions (eg. mutated genes) and the results (eg. phenotypes) are not fully established. We are using two complementary languages, statecharts and live sequence charts (LSCs) , which describe system behaviors using time-constrained logical constructs. Here we present our model of VPC fate specification using the language of LSCs and the Play-Engine tool. The Play-Engine allows data and mechanistic models to be entered by the manipulation of a graphical user interface (GUI) that resembles the pictorial models that biologists draw ( play in ). The GUI also functions as a dynamic pictorial model that depicts computer-executed simulations ( play out ). Using these tools, we have succeeded in representing various types of data commonly obtained in VPC specification experiments, the variability of biological results, cell movement events, and the interdependence of biological behavior and anatomical structure. The model is easily extendable, either by including additional VPC specification data, by linking together GUIs that represent other developmental processes, or by creating links to models in the complementary language of statecharts . The construction and analysis of our model of VPC fate specification has enhanced our understanding of this biological process in several ways. At the simplest level, we have transformed biological data and mechanisms into a database of executable statements of system behavior. Beyond this, however, our modeling has revealed inconsistencies between the "rules" and the actual data in these established papers, and has highlighted interesting gaps in our understanding of this system that we have begun to investigate experimentally.

Naaman Kam et al. (2005) International Worm Meeting "A computational model of vulval fate specification"

Hillel Kugler, E Jane Albert Hubbard, Naaman Kam, Amir Pnueli, Michael J Stern, David Harel

[International Worm Meeting, 2005]

We have developed a computational model of vulval precursor cell (VPC) fate specification. This model is capable of reproducing essentially all of the experimental observations reported in a subset of papers that form the basis of our initial understanding of this process. Since this model is mechanism-based, it is also capable of predicting outcomes of new experiments. Our results demonstrate that the development of such computational models can be helpful in integrating the complex, dynamic behaviors of biological systems. The generic nature of the tool and methodologies used for this project enables the natural extension of our model to other aspects of C. elegans biology. We used the language of LSCs and the Play-Engine tool to build a model that dynamically recapitulates the experiments presented in three seminal papers [Sternberg and Horvitz (1986); Sternberg (1988); Sternberg and Horvitz (1989)]. Our model is driven by biological mechanisms inferred from results reported in these papers as well as several more recent papers. The Play-Engine allows both experiments and mechanistic hypotheses to be entered in a user-friendly manner via the manipulation of a graphical user interface (GUI). The GUI also reflects the dynamic representation of the computer-executed simulations. In constructing the model, we represented all of the biological behaviors and phenomena referred to in these papers including developmental time, cell movements, gradient formation, cell fate assumption, variable biological outcomes, and inter- and intra-cellular signaling. The model enables the user to perform all experimental perturbations described in these papers, including cell ablations and genetic manipulations. The construction and analysis of our model of VPC fate specification enhanced our understanding of this biological process in several ways. First, we identified mechanistic gaps that have not been investigated. One such mechanism governs the movements of the VPCs that allow them to replace one another according to their developmental hierarchy. Second, constructing the model provided unexpected insights into the biology of this system, such as the dynamics and prioritization of fate assumption. These insights have instigated a number of experiments whose results will further enhance the model and our understanding of this process.

Na'aman Kam et al. (2002) European Worm Meeting "A GUI depicting VPC specification Formal Modeling, Simulation and Analysis of C. elegans Development"

Rami Marelly, David Harel, Na'aman Kam, Albert Hubbard, Irun R Cohen, Amir Pnueli, Michael J Stern

[European Worm Meeting, 2002]

Hillel Kugler et al. (2007) International Worm Meeting "Modeling C. elegans vulval cell fate specification."

Naaman Kam, David Harel, Michael J. Stern, Lara Appleby, Amir Pnueli, E. Jane Albert Hubbard, Hillel Kugler

[International Worm Meeting, 2007]

We are testing the feasibility of applying a rule-based, logic-driven approach to modeling biology based on state-of-the-art methods and tools from software and system engineering. The model is assembled from short logical statements that each describe a small part of how the system works. These descriptions of the various snippets of biology are termed scenarios. There are two sets of scenario statements in our model. One set describes the rules we have inferred about how the system works; these propel model simulation. Since these are generalized rules of behavior, the model is predictive. The second set of scenario statements describes actual experimental outcomes; these can be used to test whether the rule-based model is consistent with the data upon which it is based. An attractive feature of this approach is that scenario-based models are assembled similarly to the manner in which we build our understanding of biology from sets of experiments that address specific aspects of a biological process. The model can be built, altered, tested, and extended in a modular fashion, recapitulating the iterative processes of hypothesis testing and data analysis. We tested the feasibility of this approach by modeling a subset of the data pertaining to C. elegans vulval precursor cell (VPC) fate specification. The model covers a wide range of behaviors - anatomical, cellular and genetic. We present a rigorously tested dynamic model that is both usable and extendable. Simulations can be run and analyzed dynamically in detail under varying conditions. We first tested the model by seeing if it could reproduce all of the data of the pre-1990 data set from which many of the underlying rules were inferred. Our modeling methodology is also useful for testing competing hypotheses. Multiple hypotheses can be represented and stored within the same model by introducing the subset of variant rules that distinguish them. The variant models can then be tested for their consistency with the known set of pertinent data to help understand their differences. We applied this feature to probe the sequential versus graded signaling hypotheses given our pre-1990 data set. Systematic testing then identified the actual experimental results that were not consistent with each hypothesis. The construction and testing of our dynamic model raised important testable hypotheses. Our work demonstrates that the traditional hypothesis-experiment cycle can be supplemented by the construction and use of executable dynamic models, creating a more powerful cycle of hypothesis-modeling-testing-experiment.

Na'aman Kam et al. (2002) East Coast Worm Meeting "Formal Modeling, Simulation and Analysis of C. elegans Development"

David Harel, E Jane Albert Hubbard, Na'aman Kam, Michael J Stern, Irun R Cohen, Amir Pnueli, Rami Marelly, Hillel Kugler

[East Coast Worm Meeting, 2002]

The field of developmental genetics is entering a new phase, in which the synthesis of information from many sources will be necessary to gain a deeper understanding of how various tissues, cells, biochemical interactions and genetic networks collaborate to form a functional organism. The purpose of this project is to model and rigorously simulate and analyze a particular biological system the C. elegans egg-laying system using languages, methodologies and tools developed by computer scientists for the reliable development of highly reactive computerized systems. The model will incorporate existing anatomical, genetic and biochemical data pertaining to the development and function of (i) the gonad, (ii) the vulva, (iii) the uterine and vulval musculature, and (iv) the hermaphrodite specific neurons (HSNs). We concentrate on an object-oriented approach using the visual language of statecharts for specifying behavior, and tools such as Rhapsody for model execution and analysis. In previous work, we have successfully applied this language and tool to the biological phenomenon of T cell activation. The T cell activation model served as a feasibility test and integrated phenomena associated with cell-cycle control, cell fate, cell behavior and location. We are now in the midst of a far more ambitious effort, involving more complex biological phenomena that will incorporate additional aspects of development, including cell fate acquisition, cell migration, axon guidance, and apoptosis. In principle, our model will eventually handle virtually all aspects of development, ultimately allowing our results to be extended to and used by the entire C. elegans community, and will apply to other systems too.