As reported in the Economist:
As he told “The next 10 years”, a conference organised by Microsoft Research in Cambridge, England, Dr Harel has been working on a computer model of C. elegans. He hopes this will reveal exactly how pluripotent stem cells—those capable of becoming any sort of mature cell—decide which speciality they will take on. He thinks that a true understanding of the processes involved will be demonstrated only when it is possible to build a simulation that does exactly—but artificially—what happens in nature. With colleagues at New York University and Yale University in America, he is modelling and testing the possibilities.
Indeed, he proposes to evaluate the result using an updated version of the Turing test. This was devised by Alan Turing, an early computer scientist, to identify whether a machine is capable of thought. The original test proposes that a person be presented with a suitable interface—say, a keyboard and a screen—through which to communicate. If the operator cannot tell the difference between talking to another person through this interface and talking to a computer, then the computer can be argued to be thinking. Dr Harel’s version is a little more challenging. He wants to test whether scientists well versed in the ways of C. elegans could tell his computerised version from the real thing. So far, the distinction is obvious, but it may not always remain so.
Here is the abstract of Harel’s talk from the PDF of the MSRC conference:
A Grand Challenge for Computational Biology
Biological systems can be modeled beneficially as reactive systems, using languages and tools developed for the construction of man-made systems. The style of modeling is not necessarily aimed at seeking answers to specific questions but at understanding an entire system in detail, utilizing all that is known about the system, and using that understanding to analyze and predict behavior in silico. The style produces realistic models that capture the behavior of actual elements, giving rise to realistic, interactive and modifiable executions that reveal emergent properties. I will discuss a long-term “grand challenge” – to model a full multi-cellular organism. The C. elegans nematode worm is a good candidate; it is complex, but well-defined in terms of anatomy and genetics. The challenge is to construct a full, true-to-all-known-facts, 4-dimensional, fully animated model of the development and behavior of this worm (or of a comparable multi-cellular animal), which is easily extendable as new biological facts are discovered. [bold added]
It is well-established that computational models have proven themselves repeatedly for modelling various aspects of biological systems. What, then, is required in order to “construct a full, true-to-all-known-facts, 4-dimensional, fully animated model of the development and behavior of” C. elegans? I would like the phrase the question in the following way:
What requirements are placed on C. elegans for such a model to fit within the boundaries of computation?
This may seem like an odd way to ask a modelling question. In particular, it seems odd because it implies placing restrictions on the system under study, rather than on the model. However, it is entirely appropriate, indeed necessary, to ask such a question. This need arises because in the task proposed by Harel, the type of model has been presupposed — namely, the model must be simulable on a Turing machine. In turn, success in the task will logically depend upon whether or not all models of C. elegans are indeed simulable on a Turing machine. In turn, this places restrictions on the material system C. elegans itself: the material system must possess only properties or qualities such that all observables of the material system, and all relations between those observables, result only in models which are simulable on a Turing machine. This is a very stringent requirement to place on any material system, and is that much more striking when that material system is a biological organism. However, this is simply a logical ramification following from the precondition of computability.
In addition, this precondition may well be at odds with most interpretations of ‘systems biology’. Usually, one of the justifications for systems biology is that there are properties which appear at the level of the entire organism, but which are cannot be reduced to, nor assembled from, properties of only portions of the organism, and so systems biology is needed so it can address such properties of organisms that cannot be addressed by other methodologies. Such systemic properties are emergent, in the sense of the lack of an ability to fractionate such properties into “atomic” properties (properties of subsystems of the organism), and conversely, an inability to derive from any collection of “atomic” properties the systemic properties. However, it has already been proven  that a system which possesses only computable models cannot possess such emergent properties. Instead, such systems have a single largest model, which can be decomposed into minimal models. Thus, the properties of the largest model cannot be other than a direct sum of the properties residing in minimal models. Thus, either systems biology is incorrect that organisms possess such emergent systemic properties, or, computational biology is incorrect that it is – even in principle – capable of the degree of modelling proposed by Harel. Both cannot be the correct.
The Turing Test technique mentioned by the Economist is discussed by Harel in a 2005 paper :
Here is where validating the model can be likened to the Turing test, but with a Popperian twist: a comprehensive model of a full organisms will be deemed valid/complete/adequate if it cannot be distinguished from the real thing by an appropriate team of investigators.
Harel notes there are a number of technical difficulties to designing and implementing such a test. However, there is a more fundamental problem. The original Turing test, as well as this version, are tests of imitation or mimesis . In particular, these are tests of the imitation of only the behavior of a system by the behavior of some simulacrum. There is no requirement that the simulacrum actually model the causal entailment underpinnings of the original system. Further, since there is no logical basis to assert that similar behaviors entail similar underlying entailment structures, successfully passing a Turing test-like procedure does not entail that that the simulacrum is a model of anything.
As a result, Turing test criteria are not at all useful for validating models. Indeed, if it were useful, then arbitrarily “curve-fitting” a model to fit the data would have to be considered as good practice, since then the model would better “validate” in such a test. Of course, we do not treat models that way; instead, if adjustments to models are made, then there must be reasons which do not reside merely in the behavior of the model, but which instead reside in the inferential entailments within the model which can be shown to correlate with causal entailments in the system being modeled.
 Rosen, R. 1991. Life Itself. Columbia Univ. Press.
 Harel, D. 2005. A Turing-Like Test for Biological Modeling. Nature Biotechnology 23:495-496. PDF
 Rosen, R. 2000. Essays on Life Itself. Columbia Univ. Press.