Parts | Days | Selection | Search | Updates | Downloads | Help
BP: Fachverband Biologische Physik
BP 33: Posters - Systems Biology & Gene Expression and Signalling
BP 33.1: Poster
Tuesday, March 21, 2017, 14:00–16:00, P2-OG1
Error model estimation by maximum-likelihood methods — •Mirjam Fehling-Kaschek, Daniel Kaschek, and Jens Timmer — Physikalisches Institut, Universität Freiburg
Mathematical modeling has become an established approach in cell biology to gain information about intracellular processes. Especially for dynamic modeling, time-resolved data is required. Depending on the measurement technique, taking data points is time-consuming and expensive. Therefore, the modeler is often confronted with the problem of low number of replicates from which uncertainties need to be estimated reliably.
Error models provide a way to pool replicate measurements from different time-points and conditions to estimate the contributions from different error sources. Here, two complementary maximum-likelihood approaches to identify error model parameters, (1) from mean-variance tuples and (2) from model residuals, are implemented. Advantages and disadvantages of both approaches are discussed and usecases from different applications presented.