This post includes musings on the limitations and benefits of studying physical biology, and how it can best be approached. One of these headings I agree with, and one I do not :-).
Biophysics is dead
A quantitative physics approach to biology has been lauded in recent years as a new way of looking at complex biological systems and applying mathematical models with predictive power. However many researchers remain skeptical about the wisdom of perusing these avenues of research. Biological systems are intrinsically messy, with large amounts of noise and are notoriously difficult to quantify with any kind of accuracy. These types of systems are best treated as "black boxes", and gaining a qualitative understanding of the workings of the box are all that is required. Understanding the molecular-scale and mechanistic interactions that are occurring is not required to understand the higher-order systems level effects. Many of the large breakthroughs in cellular biology have occurred without understanding atomistic mechanisms, but by intuiting rules based on the larger-scale "observables" outside of the cellular box. Many pillars of the cellular biology field were trained before the advent of genomics and biophysics, making submerging more than toe-deep into this field a frightfully difficult endeavour.
Those who hope to define useful mathematical models of biological systems with predictive power face a conundrum. All of the components involved in a given biological process are rarely understood, or even known. Without complete knowledge of all the components and interactions involved, one cannot make a useful mode. After all, a new long non-coding RNA transcriptional factor or metabolite could be discovered which is not included in the model. If however there is a system in which all of the relevant components and their interactions are well understood, a model could be made -- yet it would serve no purpose, considering all the interactions and components are understood already.
Long live physical biology
The quantitative approach to biology has been lauded in recent years a sa new way of looking at complex biological systems and applying mathematical models with predictive power. Applying models to well-understood systems to better understand the mechanistic details and rates is certainly useful, but the true utility of physical biology comes from applying basic physical models to complex systems which we do not completely understand (black boxes). Having a toolbox of basic physical models -- like those governing diffusion, statistical mechanics, or quantum mechanics allows one to make quick testable hypotheses about the physical mechanisms behind a process, given even meager amounts of data about the inputs and outputs of the black box of interest. If a simple application of a physical model agrees with your data, then you've gained mechanistic knowledge of the process. If it doesn't agree, then you've still gained mechanistic knowledge of the process, if only knowledge of what mechanism is not behind this process.
Essentially, having an understanding of physical mechanisms involved in biology allows one to understand data and intuit meaning beyond a series of data points. If a rate of a signal transduction across a cell is many orders of magnitude faster than the Brownian diffusion time of a typical protein, then it can be understood that some type of active transport is involved. Other attributes of the rate data may be characteristic of stochastic focusing, or signal amplification, which contributes more information about the physical processes that may be involved. This approach certainly does not seek to give a complete picture of the molecular-scale interactions, but just as importantly it suggests to the enterprising researcher what is interesting and unexpected, and casts a light on what to look for.
Intuition in learning physical biology
If I've learned anything in my many years in school, it is that concept only begin to be useful once you've moved past the memorization and intricacies. However, the process of moving past this stage into a more intuitive understanding of a concept is one that cannot be rushed, and certainly cannot be easily skipped. Individuals with a hard-won intuitive understanding of a field are often very passionate about communicating the big picture, but they must be mindful not to skip the necessary (but often tedious) foundations. An easy example that anyone who has studied chemistry will immediately appreciate is that of organic chemistry. The first two courses of general and then introductory organic chemistry are incredibly taxing. The sheer number of mechanisms, concepts, and synthetic "tricks" that are introduced is intimidating and off-setting for all but the most intrepid young chemists. But by the second or third course of organic chemistry, everything begins to click. Reactions that you haven't seen before become much more memorable, because they have begun to make sense. Your mind has seen enough that it has laid a groundwork of rules and concepts which even the most complicated chemical mechanism can fit into. Eventually you find yourself not looking at carbons and oxygens but at groups that look like the want to get the hell away from that non-polar mess, or groups that look like they positively love electrons. Not thinking about the acidity or formal charge of amines, but about how happy this molecule looks, or how willing it would be to form a ring.
The point of a course in physical biology should be to instill this type of intuitive groundwork for the physical mechanisms behind many biological phenomena. A firm understanding of thermodynamics and statistical mechanics can give one an eye for enthalphic benefits of binding and their allosteric consequences, or the entropic penalties of mixing and the self-assembly of structures which often result. Understanding mass, energy and momentum transport can give similar advantages in understanding the dynamics inside of the crowded cell. Continuum mechanics can make you look at macromolecules not as abstract assemblies of amino acids or nucleotides, but as deformable chains which follow well-understood trends and laws. As described above, this type of understanding and ability to fit data to an existing underpinning of physics is incredibly useful.
But again, it takes time for such a groundwork to develop. Gaining a firm understanding of any of these topics takes much more time than a short crash-course. Simply glossing over physical concepts which learners are being introduced to for the first time cannot give the type of understanding that I am referring to. Thus, any physical biology course should be targeted at learners who already have some degree of preparatory study in physics or engineering. For these learners, the groundwork for understanding physical phenomena already exists; it just needs to be put in a biological context.
The key to accomplishing this should be repetitive application of physical models and mechanisms to real biological systems in the form of vignettes. Without needing to take the time to explain and derive the physical models, much more time can be spent on giving biological context, and linking physical models with their applications. The course should be laid out based on the physics behind the biology, rather than by traditional arrangements of cellular biology topics. The similarities of actin and cytoskeletal structures with DNA bending and compression and the potential applications of beam theory would merit their inclusion in the same section. A typical topic would be introduced by explaining its biological relevance (the assembly of virions in cells, for example). Then a simple physical model (such as the compression of a Hookian spring) would be applied, making it clear what assumptions were made and the reasons for their validity. Then the nearness of the physical fit with the observed phenomena would be compared. It would often be just as useful to show how some assumptions or physical models are poorly suited for a topic (such as simple diffusion for neural signaling), to get a feeling for what it looks like when you've oversimplified your model.