Biophysical Society Bulletin | March 2019

Career Development

This monthMolly Cule invited a guest to answer the question HowDo I Explain a Computational or Theoretical Model to Experimental Biologists?

It is not an easy job. Let me just draw your attention to one important thing that people may ignore: The model can be useful, but that is conditional on how one asks questions and formu- lates the model. First, let us revisit what a model can provide and cannot. A model is con-

of interest are time averages or dynamical measurements, and whether stochastic or deterministic methods are needed, among other considerations. Consequently, the applicability and reliability of the model depend on how closely the model resembles the original system, whether the simplifications have been reasonably made, how well the current model fits with the existing knowledge or experimental data, and whether the model is able to predict experimental measur- ables for realistic tests. In summary, if you’re an expert specializing in theory or mod- eling and simulation, let your experimental colleagues know which types of questions they can ask, as well as the power and the limitation of your specialties. If you pursue quanti- tative modeling in general, you do not only want to tell your experimental partners the model conclusions or predictions, but also want to let them be aware of the assumptions of the model, the reasoning possibly, the essentialities of methods, and even some caveats. As such, when the model conclusions or predictions appear awry, you know that you don’t have to blame the model or yourself first; you can remedy your model assumptions or settings to see if that helps. After all, all mod- els are wrong or simplified, but some are useful, and a few, hopefully, can be interesting and inspiring. Jin Yu Beijing Computational Science Research Center

ceived as a simplification of reality. We build a model based on existing knowledge and reasonable assumptions. One can use the model to draw insightful conclusions and make testable predictions. Accordingly, the accuracy of conclusions or pre- dictions depends on the reliability of the underlying model or assumptions, which are unknown or to be examined. Computational biophysics may deal with modeling and sim- ulation of bio-molecules, for example, to reveal fundamental molecular dynamics related to biological functions. The force fields that specify the forces to be implemented onto the modeled particles have been physically as well as empirically determined. That means simplifications and assumptions have been involved in formulating the molecular model, e.g., from the quantum mechanics level to classical atomic level, or to a further coarse-grained level. To obtain sensible answers regarding a particular type of model, questions need to be asked at a commensurate level. Be aware that as a more refined structural resolution is demanded, a more restricted simulation time scale can be achieved for investigating the system dynamics, under a limited amount of computational resources. That is to say, if one is interested in long-time dy- namics of the system, one has to sacrifice molecular details, and vice versa. For phenomenological models that can be flexibly formulated at any level of interest, the quality of the model relies more explicitly on the model assumptions and set-up. In quantita- tive biology or study of a complex system, data sets collected experimentally may demand quantitative modeling first, before probing the highly elusive underlying physical mecha- nisms. To formulate the model, a modeler can communicate with experimental partners on the basics of the systems first: For example, how many particles or subunits are involved, are there interactions or coupling among them, and which types, i.e., the nearest-neighbor interaction or a long range type. Then one has to make it clear whether the quantities

Cover Letters Are Annoying, but Here’s How You Write Them March 29, 2019, 2:00 PM Eastern Presented by Alaina G. Levine Work-Life Balance: It Can Be Done, You Can Have Fun (In Both Worlds) June 19, 2019, 2:00 PM Eastern Presented by Alaina G. Levine

Register Today at biophysics.org/webinars

March 2019

10

T H E N E W S L E T T E R O F T H E B I O P H Y S I C A L S O C I E T Y