Endy:Measkit draft

Back to PiP

=Ideas, Issues, & Questions Related to Measuring the Activity of a Promoter=


 * PoPS, polymerase per second, is a derivative unit based on the SI base units of Mole and Second. PoPS represents the number of polymerase molecules that pass any specific position on a nucleic acid per unit time.
 * Q. Is derivative the right word?  Compound feels like a better word.  See here.
 * Jason R. Kelly 12:38, 12 August 2008 (EDT): yeah, not clear. Seems like it is pretty much equivalent to the katal (catalytic activity, mol/s) which is a derivative unit.  I couldn't find any info about the bar to clear for being a derivative unit.  it may actually mean it's been officially "cleared" by some official SI body, so we might not want to just use it w/o looking into it more.
 * I think that derivative units are simply officially named compound units, so we want to use the word compound not derivative.
 * I do not believe that you are correct about the katal. That is a really weird unit, having to do with the amount of a catalyst needed to carry out a reaction.
 * Jason R. Kelly:Agreed. Wasn't proposing a PoPS replacement, sounds like we should go with compound.
 * We are proposing that a relative measurement of promoter activity may be useful for reducing variation (across conditions and settings) in the observed absolute activity of promoters. In practical terms we are proposing that the activity of BBa_J23101 be used as a physical reference standard to which the activity of other promoters may be compared. Instead of measuring the absolute activity of a promoter in PoPS, the activity of the to-be-measured promoter and BBa_J23101 are both measured under the same conditions and a unitless ratio is taken (i.e., PoPS/PoPS []= unitless).
 * Q. What should the resulting, unitless number be called, if anything? Standard Promoter Units, SPUs?  Or...
 * This specific question has been tricky to resolve. It may be useful (or dangerous) to make comparisons to other types of measurements, and to identify and discuss the issues which led to their success.  Here goes...
 * 'Jason R. Kelly 12:43, 12 August 2008 (EDT):The term "absolute" may make things more confusing. All SI measurements are relative, for example the mol and second from which PoPS (the unit of promoter activity) is derived are relative measurements.  So I worry about framing the discussion as absolute vs. relative.
 * OK. So, relative is a valid and safe word but we a replacement for absolute.  How about physical or molecular?

Example 1. Gel Electrophoresis.

 * Slab gel electrophoresis (PMID 11116182) is now used routinely to separate and estimate the mass or length of proteins, RNA, and DNA. Molecules within samples are separated in space and then observed.  Typically, the absolute distance that a molecule travels through a gel is not directly converted to the size of the molecule (with apologies to equipment that allows for direct model-based estimation of length or mass from distance traveled, run time length, or some other observable).  Instead, reference standards of known length or mass are used, the distance traveled by both the sample and reference are measured and compared, allowing for a relative measure to be made (thereby accounting for differences in experimental conditions).  However, since the length or mass of the reference standard is previously known the relative measure can be converted to an absolute measure of the sample such as base pairs or Daltons.
 * Jason R. Kelly: What is meant by an "absolute measure"? Maybe a better description would be "a measure unaffected by measurement conditions"?
 * There is an important difference between the activity of a promoter and the length of a fragment of DNA. The actual distance that a fragment of DNA travels on a gel may vary from one experiment to the next but the absolute underlying length of the DNA molecule being characterized does not change. In contrast, the reported absolute activity of a promoter (as measured in PoPS) may vary greatly across experimental conditions, and such variation may reflect true changes in the underlying activity of the promoter.
 * Jason R. Kelly 14:31, 12 August 2008 (EDT): I agree with this difference between DNA and promoters. The property of DNA measured in units of basepairs is unchanging based on experimental conditions.  The property of promoter activity as measured in PoPS changes depending on conditions.
 * Another point to ponder is that folks are familiar with (i.e., prefer to have knowledge of / use) the absolute length of DNA. 'Back in the day,' researchers didn't define the smallest band from a HindIII lambda DNA digest to have units of 1.0 "NEB" and then report everything in units of NEBs.  Will gene expression engineers prefer to have promoter activity described in some reference term (i.e., relative to 1.0 standard promoter units, or SPUs)?  Or, will absolute molecular activity defined in terms of well-defined base units be preferred (i.e., PoPS)?
 * Jason R. Kelly:Agree with this. Basepairs had value in that they fit into a "framework for understanding biology" that for instance DNA length as measured in meters wouldn't have fit into as easily.  Seems like you would rather have promoter activity in terms of PoPS, but here's that "technical" reference on the Ohm I promised I would dig up that talks about defining the Ohm as a relative unit equivalent to a physical object (e.g. a "NEB"):  First 10 pages of this (Reports of the Committee on Electrical Standards Appointed by the British Association for the Advancement of Science - Jenkin, Kelvin, Joule, Maxwell).
 * I'll start by noting that you don't actually state a conclusion or make a point here, you instead ask me to read a reference. Now having re-read this reference here are some questions:
 * Q1. Was the exact value of the Ohm defined under standard conditions of measurement (e.g., as was the prototypical meter)?
 * Yes.
 * Q2. Is it true that there is no intrinsic value in naming a unit (e.g., THE OHM) other than convenience?
 * No that is not true. In this case the Ohm was equal to the resistance of a physical object it couldn't have been described in base units.  They had to name it something.
 * Q3. How is the world of standards different today compared to the 1860s? What is the significance of the fact that we now have well-established base standards for time and number?
 * Not sure. One big difference is the quality of reference standards has improved significantly and most are tied to universal constants (e.g. speed of light) rather than to physical objects (with the exception of the kg which is still a hunk of metal).
 * Q4. Was not the physical reference object for the Ohm itself selected and manufactured with the goal that it, ideally, provide a prescribed amount of resistance?
 * Yes, that is correct. They chose to target it to a value that made calculations easier (I think) when using base units.  However the Ohm was not equal to some value in base units it was equal to resistance of a physical object.
 * It's interesting to consider the history of units of time: for example, "The Egyptians subdivided daytime and nighttime into twelve hours each since at least 2000 BC, hence their hours varied seasonally." Later, with the development of pendulum-based clocks, there was an transition from apparent time to observational or mean time.  From there to the rotation of the earth, to it's rotation w/in a particular epoch, to atomic clocks under defined standard conditions.
 * As the second transitions from apparent to mean, other units of measure can be developed based on the definition of a second.
 * Imagine how things would be different if we used apparent time today. In the northern hemisphere winter, would meetings would last "longer" at night versus the day?  Or, would daytime meetings be scheduled for a longer block of time? 

Example 2. Length.

 * "The metre or meter is a measure of length." From this same reference, we find that on September 28 1889, the "first General Conference on Weights and Measures (CGPM) defines the length as the distance between two lines on a standard bar of an alloy of platinum with ten percent iridium, measured at the melting point of ice."  Later, on October 6 1927, "the seventh CGPM adjusts the definition of the length to be the distance, at 0 °C, between the axes of the two central lines marked on the prototype bar of platinum-iridium, this bar being subject to one standard atmosphere of pressure and supported on two cylinders of at least one centimetre diameter, symmetrically placed in the same horizontal plane at a distance of 571 millimetres from each other."  Most recently, on October 21 1983, "the seventeenth CGPM defines the length as equal to the distance travelled by light in vacuum during a time interval of 1⁄299,792,458 of a second."
 * So, if something heats up and becomes longer, it's length as defined in meters will be said to increase, because the definition of a meter is under some defined standard conditions.
 * Jason R. Kelly:Yes, that's right.

Big Question 1.

 * A Prototype or Reference Promoter? Should we have a reference promoter that embodies a standard, to the best of our ability? We can define promoter activity in terms of promoter clearance rate and more generically in terms of PoPS, both of which can be related to base SI units.  Thus, one path forward would be to define 1.0 Standard Promoter Units to be embodied by a promoter whose activity is closest to 1.0 PoPS under some defined conditions.
 * Given that the activity of a promoter can be defined via a compound unit derived from base units that themselves already have well defined physical reference standards, it seems like a potential trap to attempt to define a new "Standard Promoter Unit" based on the activity of some new physical reference standard. Instead, we could do two things.  First, we could declare that 1.0 PoPS under some defined conditions is 1.0 standard promoter unit (SPU).  Second, people could simply make reference collections that are known to produce certain PoPS levels under various conditions.  For convenience, over time, researchers could perhaps develop ever-improving promoters that are evermore closer to producing exactly 1.000000000 PoPS, and so on (this would be similar to how ever better DNA ladders are developed, with the 2-log ladder being pretty tasty!). If the SPU becomes a useful shorthand then terrific, otherwise no worries.
 * Jason R. Kelly: see note on this ref above. First 10 pages of this (Reports of the Committee on Electrical Standards Appointed by the British Association for the Advancement of Science - Jenkin, Kelvin, Joule, Maxwell).
 * Please see my response and questions above.
 * Jason R. Kelly: I agree with much of what you say above, I provided the reference to site an example where previous engineers were worried about creating a scenario that required making an ever improving reference standard (I am not yet sure if it is as big of a deal in our scenario as it was with the Ohm, thinking about your Qs above). However, independent of that point, I am trying to reconcile the suggestion to not use SPUs with the fact that getting SPU = 0.2 across 8 different measurement conditions when measuring J23150 is clearly of value to an experimentalist trying to choose among sets of promoters. To be more clear:
 * 100 independent labs each characterize a different promoter under different conditions, and report the promoter activity relative to a reference standard (e.g. in SPUs).
 * I measure a promoter X in my lab relative to the same reference standard (e.g. in SPUs) and upon testing in my engineered system find that it is 2X as strong as i would like.
 * I look to the list of 100 promoters and find one (or a few) that are characterized as half as strong as the promoter I tested. I plug one of them in to my engineered system and find that it is also half as strong promoter X under my experimental conditions .
 * If the 100 labs has measured their promoters in PoPS the measurements would be incomparable (e.g. unable to rank order the promoters) as they were measured under different conditions
 * Jason R. Kelly:So it might be the case that we need to use a unit like SPU even though PoPS fits better into a "framework for understanding biology." Using PoPS as the unit would likely result in incompatible units due to the challenge of replicating standard conditions in different labs.  So promoter activity measured in PoPS would be incomparable unless measured all in the same lab.

Question

 * Seems like there are three options:
 * Each option consists of two parts:measuring a promoter and using a promoter measured by someone else.
 * Option 1: PoPS no ref standard
 * measuring a promoter:When measuring a promoter X a researcher should measure the activity of promoter X in PoPS. So promoter activity would be reported in PoPS along with a description of the measurement conditions used.
 * using a promoter measured by someone else:Assuming everyone didn't use identical conditions then you would need to determine the conversion between the conditions each promoter was measured under and your measurement conditions in order to use any given promoter. This approach seems infeasible with current technology.
 * Option 2: PoPS w/ ref standard
 * measuring a promoter: When measuring a promoter X a researcher should measure both the activity of promoter X as well as a the activity of a reference standard promoter in PoPS. If they want to report the standard activity of promoter X they should use the activity of the reference standard promoter in their conditions as a conversion factor to convert to the predicted activity of promoter X in PoPS under standard operating conditions (SOC).  So standard activitywould be reported in PoPS under SOC.
 * using a promoter measured by someone else: The activity of promoters that you want to use off the shelf should be listed as PoPS under SOC, so you need to first determine the conversion from SOC to your experimental conditions - this could be done by measuring the activity of the reference standard promoter (or any number of promoters that had been measured under SOC) under your experimental conditions to determine a PoPS_SOC to PoPS_expConditions conversion function. You then use then apply this conversion function to all the promoters on the shelf to find the right one based on its predicted PoPS_expConditions for your conditions.  This approach is OK, though there may be some confusion in calling everything PoPS
 * Option 3: SPUs
 * measuring a promoter: the activity of a reference standard promoter and then report the activity of promoter X in SPUs as a relative measure of PoPS_promoterX / PoPS_ref_standard_promoter. So standard activity would be reported in SPUs under SOC.
 * using a promoter measured by someone else: The activity of promoters that you want to use off the shelf should be listed as SPUs under SOC, so you need to first determine the conversion from SPU to PoPS in your experimental conditions - this could be done by measuring the activity of the reference standard promoter (or any number of promoters that had been measured in SPUs) under your experimental conditions to determine a SPU to PoPS_expConditions conversion function. You then use then apply this conversion function to all the promoters on the shelf to find the right one based on its predicted PoPS_expConditions for your conditions.  This approach is completely identical to option 2 it just gives a different unit to PoPS measured under SOC (SPUs), this may make things more or less confusing im not sure.


 * Jason R. Kelly: I think we need to go with option 2 or 3 (e.g. important to report promoter activities as if they had all been measured under SOC by using a ref standard). One reason to use SPUs is that it would make clear when a person was reporting a standard activity of a promoter (e.g. promoter activity under SOC) vs. reporting the activity a promoter under some random experimental conditions (reported as PoPS).  If we don't use SPUs then people need to be good about specifying whether they converted promoter activity to standard activity or not since everything will be in PoPS.

Possible Big Conclusions.

 * We focus the paper as highlighting the utility of making relative measurements.
 * Yes
 * We discuss that 1.0 PoPS produced under standard conditions is 1.0 SPUs.
 * This could be restated as we should make a reference standard promoter with PoPS = 1.0 (since I think we need a ref standard promoter, but may not need SPUs). I think it is premature to expend effort to make a 1.0 PoPS promoter, activity varies too much with conditions and the one conditions we set it to 1.0 PoPS in would be useless to most other people.
 * Is there something to say about the possible utility of the SPU? (see 4th point below)
 * There is clear need to have a standard activity that is the predicted or measured activity under standard operating conditions (SOC). Whether this is called PoPS_SOC or SPU is a semantics / community question.
 * We focus attention on the utility of (a) making improved promoter reference collections, and (b) making more accurate measurements of promoter activity.
 * If the suggestion implicit in promoter reference collections is that they act like NEB ladders I think that is wrong. Your point about promoter activity being different from DNA length is correct, promoter activity varies too much with measurement conditions to expect to have a Promoter that has PoPS = 1 across many conditions.  Sorry if I'm misinterpreting this.
 * making more accurate promoter meas is always good, i don't think it needs to be a big point in the paper though
 * We note the uncertainty regarding an exact or relative unit of measure for promoter activity. E.g., it could be that an exact unit of measure would require too much correction across changing conditions.  Or, it could be that an exact unit of measure would allow for precision of engineering, and that such precision will be somehow essential.
 * '''I'm a little confused here. I think when you say "exact unit of measure" you are describing something akin to my option 1 above?  And I think for our purposes that important aspect of a "relative measure" is that it lets you convert back to some set of conditions where all promoters are measured relative to each other (e.g. option 2 or 3).

Jason's Summary

 * PoPS is a unit for measuring or quantifying promoter activity (and also other objects)
 * PoPS is a derived or compound unit, and thus does not need a reference standard of its own (for now)
 * PoPS is a 'ratio scale' as defined in the 4 classes of "On the theory of scales of measurement" (Stevens, Science, 1946) since it has an absolute zero.


 * If we want to enable the measurement of promoter activity in PoPS then we need to provide:
 * At least one measurement procedure
 * A 'principle of correlation' (e.g., a function that relates the measurement of the procedure to PoPS)


 * The measurement instrument is difficult to control reliably
 * GFP coding sequence, GFP, fluorescence detection device


 * The measurement conditions are also difficult to control reliably
 * Cells, media, polymerase per cell, et cetera


 * Instruments could be calibrated across labs (would also require controls for things like plasmid copy number since that is part of the "instrument" as well. Which would be difficult to control for.)
 * For example, distribute purified GFP, beads, et cetera


 * Conditions are more challenging (and sometimes impossible) to control.
 * For example, it is now premature to define a standard strain or media for many biotechnology applications
 * As a second example, there are practical challenges of diversity of lab equipment and manufacture
 * A traditional solution is to develop a model for how the quantity being measured varies across conditions (e.g., coefficients of expansion) or to hold conditions constant (e.g., delimited temperature ranges).


 * Thus, if you want to report the activity of a promoter in PoPS by our measurement procedure with either a new instrument or under novel conditions (note above, all conditions may be novel) then you need to both:
 * Calibrate any new measurement instrument using an instrument calibration standard (e.g., purified GFP, possibly beads).
 * Measure a reference promoter (any promoter measured previously as described here) under current conditions in order to establish an empirical model that accounts for the effect of differences in conditions on activity.
 * NOTE. If only the reference standard is measured, without calibration of the instrument, then differences in conditions (but not instruments) can be accounted for; this allows for measurement of relative, but not actual, promoter activities under those measurement conditions.
 * Interestingly, a single measure of relative promoter activity may be a more useful or less confusing than reporting the actual activity of both the sample promoter and the reference promoters.
 * Stated differently, there are three options for measuring a new promoter.
 * First, calibrate the instrument (e.g., purified GFP) and control for variation in conditions (e.g., ref promoter) thereby allowing for reporting of both promoters in both relative and actual units.
 * Second, do not perform an instrument calibration, but only perform the conditions control, thereby only allowing for a relative measurement.
 * Third, only perform the instrument calibration thereby allowing for reporting of the activity of the sample promoter in actual units (but only under those specific conditions).


 * Note, stating a seemingly obvious but subtle point, accounting for conditions only matters if one is interested in the activity of a promoter across different conditions.

Starting Comments

 * How should we measure promoters under varying measurement conditions?
 * There is likely an analogy here to measuring length using a meter stick with a different (or the same) coefficient of expansion (compared to what, the thing being measured?). An open question is whether J23101 when measured under any conditions should always = 1 SPU (even if the amount of PoPS produced is changing).  As I understand this is not the way that it would be done when measuring length - if you measure a meter stick at a temperature where it expands 8X then it is 8 meters long, not 1 meter.  However, there seems to still be value in treating J23101 as = 1 SPU across different measurement approaches in that it would allow you to rank order promoters measured in different environments by different researchers.  So not sure how to reconcile these two different points.
 * How should the reference standard and unit be defined?
 * For instance, we could define the unit of the SPU as the # of PoPS from a reference standard promoter (i.e. J23101) under some standard operating conditions.
 * We could alternatively define the SPU as = 1 PoPS.
 * I think that the first option is likely to be the right one, but I think it could be influenced by the answer to the first question above.

===Other places where parts characteristics are very dependent on conditions or where relative measurement across many conditions has proven more valuable than a single 'absolute' measurement in one condition===
 * Comparative anatomy?
 * "Today comparative anatomy has reached such a point of perfection that, after inspecting a single bone, one can often determine the class, and sometimes even the genus of the animal to which it belonged, above all if that bone belonged to the head or the limbs. ... This is because the number, direction, and shape of the bones that compose each part of an animal's body are always in a necessary relation to all the other parts, in such a way that - up to a point - one can infer the whole from any one of them and vice versa."