Let’s continue now with standard biological parts.
In principle, the standardization of biological parts -i.e., gene constituents, genes, small gene networks- should allow us to simplify the process of transferring the characteristics of one biological system to another, very much like installing a new software in our computers.
Nevertheless, the issue remains that, the behavior of one biological part in a particular organism cannot be taken as an accurate prediction of how it will work other organisms, even with the most thoroughly characterized parts.
To overcome this major issue, some insights from current research can be taken into account.
We know that there will be variability, but having a measurement of how much variability is there, enables us to make experiments to figure out design principles to build more robust systems.
For example, in Toni and Tidor, (2013)
an interesting model is proposed to account for extrinsic and intrinsic variability in genetic circuits. To model intrinsic variability -that which is due to the stochasticity of the collisions between molecules-, the authors used the omega-expansion of the probabilistic master equation and to model extrinsic variability -that which is due to variations on the upstream components of the system, such as number of ribosomes, RNA polymerases, etc.- the authors didn’t use fixed parameters, but values that follow a probability distribution. After applying their method to analyze different small circuit topologies, the authors found that negative feedback, high protein cooperativity and imperfect complementarity between miRNA and mRNA, along with correlated expresion between these two, help to decrease the variability of their circuits when compared to a circuit without such designs.
Identify sources of variability
There are some experimental conditions that are widely acknowledged as variability sources when trying to transform an organism, ranging from the strain of the organisms we’re using to the growth media we’re feeding them.
Cardinale and Atkin, (2012)
, offer a comprehensive account of sources of context dependencies. The authors divide the contexts dependencies in compositional, host-borne and environmental-borne and make a special emphasis on how synthetic systems can sequester the natural machinery of the host cell -a machinery that took a millions of years of evolution to be put together and fine-tuned- and change the energetic balance of the cell, ultimately leading to cell instability.
Recently, Pasotti, et al., (2013)
, have found that different circuit topologies can affect the measurements of promoter activities. As systems get more complex, the variability on the function of its components also increases.
A proper characterization and circuit implementation should take these context dependencies into account. Which bring us to our next point.
Characterization and systematization
When talking about characterization in Synthetic Biology, the reference par excellence is the Canton, Labno and Endy, 2008
paper, where they introduce the concept of a datasheet for biological parts, similar to the datasheet of mechanical parts. This has developed to an international effort to characterize biological parts and devices, with centers such as CSYNSBI
on the lead and many laboratories committed to the task.
Furthermore, the information that result from this characterization efforts will not remain as raw, unexplorable data; computational tools are being developed to ease the extraction of meaning out of such information.
Such tools are in continuous development and in that process, as Lux et al., (2012)
state, value is being generated by ” the formalization of the design rules that determine the complex relationships between genotype and phenotype”.
…and some final remarks
Variability due to context dependencies is a formidable challenge. I cannot tell how many times I’ve heard or read pessimistic opinions about Synthetic Biology that focus on the variability of biological systems; some of them would go as far as to say that the field is ultimately condemned because of this.
Of course, it’s not like we should pretend that variability is not there and go do experiments: my point is that stating that variability is an important issue with so much pessimism is totally a different thing than approaching it as a challenge.
There are several leads to follow to measure and get closer to gain control over such variability.