(411c) Labml: Automatically Connecting Experiments And Simulations

Authors: 
Denney, W. S., University of Pennsylvania
Diamond, S. L., University of Pennsylvania


Large scale systems modeling requires a tremendous number of derived kinetic parameters, large experimental data sets, and complex model structures. Although excellent databases exist for many parameters, there is no organized framework for connecting the experimental parameters and the simulated results. Labs that are capable of generating high throughput data would benefit from direct crosstalk between experimentation into high throughput simulation. We are developing a generalized laboratory markup language (LabML) that allows experiments and simulations to run under the same protocol. It combines generalized laboratory unit operations to aspirate, dispense, separate, incubate, and measure during an experiment and it allows for simulation of each experimental. We have implemented an initial file format in our lab to provide communication between high throughput screeners and computer modelers. This XML format is used to completely describe the experiment and then it is able to be merged with an SBML model that will allow prediction of the results of the experiment. An essential feature of LabML is data provenance, providing detailed information about the history and manipulations of a data set. The LabML file is permanently stored in a database so that experiments may be compared and simulated with new models as they are developed. Ultimately, experimental results (e.g. concentrations over time) are used to generate the derived parameters (e.g. kinetic constants) necessary for building a computational model.