Endy:Measkit PLO/v2

Overall

Important points

1. PoPS is an appropriate unit for promoter activity
2. Promoter activity is sensitive to experimental conditions
3. Promoter activity may (or may not) vary equivalently within a set of distinct promoters as they are placed in different conditions
4. Relative promoter activity may be reported in Reference Promoter Units (RPUs) by normalizing to a reference standard, and may allow for the specification of a range of promoters and a range of conditions across which relative promoter activity is predictable. (key point)

Things we'd like to leave out

1. Calibration of instruments (e.g., using purified GFP of known amounts).
2. Reporting our measurements in PoPS on axis in figures. We didn't seriously determine the values of parameters in our GFP->PoPS model for any of the different conditions. We can say that this is significant work, include paragraph in discussion making this point. Also say we need to get PoPS landscape at some point in the discussion.
   • Need to have a paragraph describing why PoPS is hard to measure. It may be useful to include a sample calculation relating PoPS to RPUs in order to demonstrate the different constants, etc.
3. RBS measurements - The only RBS measurements were taken in the Endy lab and we lack the significant multi-condition data and the lab-lab data that the promoters have. Also it requires a separate run through of the model and presents a more challenging set of issues with canceling some parameters since different RBS = different mRNAs. My main concern is that it distracts from the overall story. We could mention, but i'd rather see it in the Supplementary materials with a mention in the discussion?

P-level outline

Introduction

1. Engineering many-component systems is made easier by developing collections of standard parts that can be reused.
2. It is easier still to predict the behavior of engineered biological systems assembled from standard parts if the component parts themselves were well characterized. (i.e., reusability is enabled by predictability is enabled by characterization)
   • be scholarly here, making reference to current parts collections, and lack of characterization thereof.
     • For example the registry of biological parts contains X promoters and Y RBSs Z of which have been measured.
3. Measurement of physical objects is well understood and has been successfully developed and applied in other domains (e.g., principle of correlation, et cetera). A couple lessons:
   1. In order to make a measurement we need (1) a method of measurement and (2) a principle of correlation. For example... in measuring length...
   2. In order to control for the effect of conditions on the object being measured, we can (a) hold the conditions constant, (b) measure objects that are insensitive to conditions (c) Specify models that predict activity across different conditions.
4. Measuring biological parts consistently has proven challenging, and may be unlike past experiences
   • It is unrealistic to expect that researchers will be able to use standard conditions in the short term. This is both due to practical constraints of replicating the exact media, cells, growth conditions, etc, in independent laboratories as well as due to engineering constrains where different applications simply require different operating conditions (much as cars must drive in different temperature and weather).
   • The property we are measuring, promoter activity, is sensitive to experimental conditions.
   • Our models for promoter activity in different conditions are incomplete / non-existent.
   • "There is no such thing as a standard component, because even a standard component works differently depending on the environment." -- New York Times, Tuesday, Jan 17, 2006, Custom-Made Microbes, at Your Service
5. Biological part measurement is not as hopeless as it looks, we can address the challenge of part performance varying across conditions by measuring relative promoter activity to a reference standard.
   • Cite example of measurement of spines as using a relative unit.
   • Over time the empirical data collected across many conditions will begin to remedy our lack of theoretical(?) models of promoter activity under different conditions.
6. Thus, to try these ideas out / begin to make progress, we designed a reference standard for promoters and developed models, that taken together allow for (accounting of some sorts of variation).
   • We did the following (i) used a reference standard to evaluate variation in conditions and instruments ourselves, and (ii) distributed a reference kit to validate the approach across multiple labs. Taken together we demonstrated the utility of the reference standards (prefatory summary here).

Results

• By mindful of bridging to the intro better.. we wanted to measure PoPS...

1. In order to measure promoter activity we needed a principle of correlation relating the observed property (GFP synthesis) to the property we are trying to measure (PoPS). To establish this relationship we made use of a mechanistic model. We used this model to demonstrate that by measuring relative promoter activity (in units of SPUs) cancel out many of the aspects of measurement conditions that are likely to effect promoter activity (e.g. GFP maturation rate).
   • note that other models would be derived for other measurement methods
   • show the figure with sensitivity analysis of model parameters.
2. Varying conditions produce different levels of observed GFP synthesis
   • Some of these conditions are likely to change the activity of the promoter itself (e.g. more polymerases in different strains), others change the parameters of our model relating GFP to PoPS (e.g. copy # of the plasmid)
   • Note that parameterizing the model for GFP->PoPS for each set of conditions is challenging and we did not do it here. As a result we can't report results in PoPS (for instance, the variation in GFP synthesis levels could be entirely due to copy number changes under each condition - we don't know).
   • Although the observed measurement (GFP synthesis) varies with conditions, we expected from the model that by reporting a relative promoter activity many of the sources of variability will fall out.
3. Relative promoter activity holds constant across several conditions
   • Again, we expect this based on our model
   • Show the plot comparing absolute measures of a single promoter to relative measures of the same promoter across 7 conditions
   • These results suggest that we might be able to colelct a set of promoters and a set of conditions where relative promoter activity (SPUs) is predictable.
4. We tested the practical value of SPU measurements by measuring 4 promoters across 6 independent labs
   • Show the figure of inter-lab measurement
   • The labs were able to measure equivalent relative promoter activities supporting the value of SPUs as a measurement unit
   • Be sure to note that the data from these labs would not be comparable without some difficulty (e.g. not a good GFP reference standard for FACS). this may not be a huge point

Discussion

1. Summary of what we did.
   • Demonstrated that absolute promoter activity varies across conditions
   • Demonstrated that relative promoter activity is constant across some range of conditions
   • Demonstrated that we could measure equivalent relative activity across labs that were using different measurement instruments (that would otherwise have provided incomparable measurements).
2. We can imagine a future engineering framework based on relative promoter activities rather than PoPS.
   • we collect the sets of conditions where promoters perform stably relative to each other
   • there may be sub-categories (e.g. stationary phase promoter performance may be the same across many different media conditions, but different from exponential phase).
   • We build a theoretical framework that models system performance based on relative activities rather than absolute activities.
     • For instance, we might expect that our system will experience a change in growth rate that might effect absolute promoter activity. Rather than keeping absolute activity constant, we might just design systems that function independent of this. Remind that natural biological systems, such as human anatomy, seem to similarly be based on constant ratios rather than constant absolute lengths.
3. Although relative promoter activity is useful, "absolute" promoter activity will e critical in some cases (e.g. strong off needed)
   • To support this we can begin to develop the PoPS-conditions landscape.
   • With a good enough PoPS-map we may be able to eventually predict the performance of new conditions a priori.
4. Future Work
   • Establish a reference set of promoters and an allowable deviation from known values for a set of conditions to be considered “approved”.
   • Evaluate many new conditions to see how many are “approved”
   • Evaluate sensitivity to the measurement approach
     • we varied some parts of the measurement instrument (for instance, plasmid copy number), but should change the FP as well
     • Insulated promoters – promoters with activities that are insulated from the surrounding sequence will be better suited to reliable measurement as they will be less sensitive to the measurement instrument used.
   • Improved measurement instruments – e.g. moving to the chromosome, using more sensitive reporters for promoter activity, using multiple reporters to better control for variation in experimental conditions.
5. Distribution, use, and improvement of kit
   • Reference standards only useful if they are adopted by a community. So need to make them readily available and easy to use…

Other

1. Use a consistent vocabulary for terms from measurement theory
   • Should probably use the Hand vocabulary as it is the most current – and in the intro he is pretty complete about providing different words.
2. Is it worth having a diagram showing what parts of the system are the measurement instrument and what parts are the measurement conditions?
   • Consider for introduction