Team:Imperial/InterLab Study


Imperial iGEM 2014

Interlab Study


The characterisation of standard biological parts is a fundamental principle of synthetic biology. In the interlab study iGEM teams from around the world measure the fluorescence of the same three GFP-coding devices after a set time of growth. This provides a large dataset to analyse variation between labs. Two devices were cloned ourselves as specified, and one was provided by the registry. Using both a plate reader and flow cytometer fluorescence measurements were obtained in absolute units in terms of moles of sodium fluorescein or relative to standard beads. In addition to showing increased fluorescence in the stronger promoter as expected, it was also seen that for the higher copy plasmid the cells appeared to be burdened. Finally, the flow-cytomoter data was used to explore cell to cell variation as part of the extra-credit assignment.

Key Achievements

  • The two required devices were cloned from the specified parts in the registry
  • Fluorescence measurements were obtained from a plate reader and flow cytometer and corresponded with expected resuults.
  • Fluorescence was expressed in absolute units of moles of fluorescein or relative to standard beads
  • Cell-to-cell variation of the devices was explored using the flow-cytometer data


This year, iGEM invited all the competing teams around the world to measure fluorescence from the same three genetic devices for GFP expression. This exciting experiment will allow the robustness of parts against a variety of different protocols and lab techniques to be assessed. Since the devices are the same, data should match between different labs. In synthetic biology, it is important that part characterisation is be consistent between different labs to be able create well-defined standard parts. We are excited to contribute to this experiment and look forward to seeing the results of any subsequent analysis of data at the iGEM jamboree!

Section I: Provenance and Release

Who did the actual work to acquire these measurements?

Xenia Spencer-Milnes

What other people should be credited for these measurements?

Catherine Ainsworth

On what dates were the protocols run and the measurements taken?

Cloning of constructs was confirmed by 5th September 2014. Transformant plates were from 10-13th September 2014. 16th September 2014 - all samples on flow cytometer and plate reader apart from additional fluorescein standard curve which was measured on 19th September 2014

Do all persons involved consent to the inclusion of this data in publications derived from the iGEM interlab study


Section II: Protocol

What protocol did you use to prepare samples for measurement?

The two construct built in house were cloned by biobrick cloning to insert E0240 as a back insert into the promoter with backbone. This resulted in the assembly scar sequence TACTAGAG between the promoter and the RBS. Assembly was confirmed by sequencing (Source Bioscience). We used the K823012 part for the J23115 promoter which contains the two mismatched basepairs compared to the original J23115 sequence. 2 mL of LB plus appropriate antibiotic (kanamycin final concentration 50 ug/ml, chloramphenicol 34 ug/ml) in a 14 ml polypropylene round-bottom tube (BD Falcon) was inoculated with a single colony from E.coli DH10B transformant plates. For each construct (labelled ‘A’, ‘B’, ‘C’ and ‘N’ for existing J23101 device [pSB3K3], our cloned J23101 device [pSB1C3], our cloned J23115 device [pSB1C3] and E0240 [pSB1C3] without promoter respectively), three colonies (labelled ‘1’, ‘2’, ‘3’) were picked and grown separately. Cultures were grown in a 37degC growth room at ~230 rpm (VWR Standard analog shaker speed setting 7). Additionally, a non-inoculated LB+antibiotic (chloramphenicol only) tube was grown alongside as a contamination control (labelled ‘O’). Extra media + antibiotic (for chloramphenicol only) was also kept aside for background controls.

What sort of instrument did you use to acquire measurements?

Figure 1. The flow cytometer system used in the characterisation experiment.
  • What is the model and manufacturer?
    • Plate reader: (BioTEX Synergy HT) and accompanying software (KC4 v3.2).
    • Flow cytometer: FACSCan with a BD FACS Flow Supply System and an Automated Multiwell Sampler (AMS) (Cytek Development) as pictured below, controlled by BD CellQuest™ software on a Mac OS 9.2 computer (figure 1).
  • How is it configured for your measurements?
    • Plate reader: this was used to measure absorbance at 600 nm, and fluorescence (excitation 485 nm, emission 528 nm, sensitivity 55, optics from top of plate, filter set 1).
    • Flow cytometer: this was used to measure fluorescence of individual cells. Each read was terminated at either 45s or 200,000 total events. Laser detection channel FL1 (488 nm) was used with sensitivity at 600 V and optical filter 530/30. Forward-scatter (FSC) detector setting E01, side-scatter (SSC) setting 375. All samples were run on LOW flow setting. The appropriate sensitivity voltage was determined using test wells to confirm that fluorescence peaks across both the high and low strength promoters could be seen clearly.

What protocol did you use to take measurements?

Plate Reader

Preparation of samples were as follows: 100 ul cells from each culture, overnight control and media control were transferred in triplicate to a black, flat-bottomed, 96-well Greiner plate on ice. Initially, a sodium fluorescein control was added to the plate in duplicate, at a concentration of 500 ng/ml and final volume 100 ul, by serial dilution in Dulbecco’s Phosphate Buffered Saline pH 7.37 (DPBS) from a 500 ug/ml stock. However, it was decided the readings from this could be improved as a control and a separate control plate was set up at a later date containing varying concentrations (500, 375, 250, 125, 50, 25, 10, 5, 0 ng/ml and PBS only) in triplicate at a final volume of 100 ul. This was set up by pipetting in serial dilution from 3x 1/100 (for the &gt 125 ng/ml concentrations) and 3x 1/1000 (for the &lt 50 ng/ml concentrations) aliquots of 500 ng/ml stock. Confirming a linear relationship between fluorescence of sodium fluorescein at different concentrations and calculating the conversion factor as an average across the different concentrations will provide a more accurate control to convert to absolute fluorescence. The plate was read in the plate reader using the settings described previously. Data was transferred to Excel (Microsoft Office 2007) for analysis.

Flow Cytometer

Preparation of samples was as follows: Each culture or control was transferred to the plate in triplicate (labelled ‘I’, ‘II’, ‘III’). In a 96-well COSTAR plate, A, B, C and N cultures were pipetted out to a 1/100 concentration using a serial dilution from 1/10, with a total volume of 200 ul per well. Test wells were used to determine the appropriate voltage sensitivity for the the fluorescence read (data is not included). 200 ul of each media and overnight incubation control was added to the plate in triplicate. A well for the standard beads was prepared initially with just 190 ul dH20. The plate layout is shown in figure 2.

General use of the flow cytometer followed the Standard Operating Procedure (SOP) document specific for the machine, as written by the lab. Settings are as described previously. The plate was loaded into the AMS, and each well was measured for either 45s or up to 200,000 events. Before the run of the samples, wells were pipetted up and down so ensure cells were suspended in solution. Beads (10 ul) were only added to the well (190 ul dH20) just before the read took place, and with vigorous pipetting to ensure they were evenly distributed in the well. Data was analysed using Cyflogic v.1.2.1. then Excel (Microsoft Office 2007).

1 2 3 4 5 6 7 8 9 10 11 12
E A1-I A2-I A3-I B1-I B2-I B3-I C1-I C2-I C3-I N1-I N2-I N3-I
Figure 2: Layout of sample plate for flow cytometer to characterise fluorescence of three different constructs expressing GFP.
  • Test wells - to determine appropriate voltage (sensitivity).
  • Excluded - voltage was changed while sample was running therefore results are invalid ‘
  • ‘A’ wells - Existing J23101 device, cultures ‘1’, ‘2’, ‘3’, sampled ‘I’, ‘II’, ‘III’
  • ‘B’ wells - Cloned J23101 device, cultures ‘1’, ‘2’, ‘3’, sampled ‘I’, ‘II’, ‘III’
  • ‘C’ wells - Cloned J23115 device, cultures ‘1’, ‘2’, ‘3’, sampled ‘I’, ‘II’, ‘III’
  • ‘N’ wells - negative control device (no promoter), cultures ‘1’, ‘2’, ‘3’, sampled ‘I’, ‘II’, ‘III’
  • ‘M’ wells - Media plus antibiotic (chloramphenicol only) background control, sampled ‘I’, ‘II’, ‘III’
  • ‘O’ wells - Non-inoculated culture (chloramphenicol) contamination control, sampled ‘I’, ‘II’, ‘III’
  • Beads - Sphero™ AccuCount Fluorescent particles 5.1 uM, ~1E6 particles/ml as reference for absolute fluorescence. 10 ul in 190 ul dH20.

What method is used to determine whether to include or exclude each sample from the data set?

Plate reader: all samples were included in the data set.

Flow cytometer: reads were gated based on side scatter (SSC) to select cells and exclude cell debris, and based on reducing the tails of the distribution of the SSC histogram (e.g. see appendix figures 2-5.1, 2-5.2 and 2-5.3). SSC gating was the same for every sample. Previous experiments in the lab have suggested forward scatter has no relationship with cell population, and this can also be seen qualitatively in our results where the events are evenly distributed for forward scatter (e.g. appendix figures 2-5.1), so this was not gated. Fluorescence was also gated since previous experiments in the lab have suggested that the population at very low or no fluorescence is cell debris rather than non-expressing cells. These gates were adjusted manually for each population to include the main cluster of cells. The resulting gating gives a log-normal distribution, hence the geometric mean is calculated and this value is used as the fluorescence value for that well. For the negative controls, only SSC gating was included since the population with cell debris overlapped in the fluorescence. The beads for fluorescence were gated on both side-scatter and fluorescence to obtain the correct population, based on previous experience, and that they should be a tight group with high SSC and FSC (figure 3A,3B,3C). Note from the bead graphs it can be seen that the water used to dilute the beads was likely contaminated. However, the population of beads is still clearly visible. Additionally, one replicate for culture A1 was not included in the analysis because voltage sensitivity was adjusted whilst the sample was running. Examples of the gating for each construct and the negative construct are shown in the appendix.

Figure 3A: Dot plot showing tight population of standard beads (boxed). Remaining events are thought to be potential contamination from diluent water.
Figure 3B: Gating of standard beads by side scatter (SSC) and fluorescence (FL1) showing tight resolution population
Figure 3C: Histogram of gated standard beads showing high resolution peak. The geometric mean fluorescence is 1391.51.

What exactly were the controls that you used?

A negative control for background cell fluorescence was included as cells containing the device E0240 but without a promoter, to mimic burden of the promoter. The LB media and a non-inoculated overnight culture (using antibiotics chloramphenicol only) were used to control for media-only background, identify any possible contamination and see the debris pattern. In order to obtain absolute values for fluorescence, set standards were measured alongside the samples.

For the flow cytometer: Sphero™ AccuCount Fluorescent particles 5.1 uM, ~1E6 particles/ml. 10 ul were added to 190 ul dH20.

Figure 4: Confirmation of linear relationship between sodium fluorescein concentration and measured fluorescence on a BioTEX Synergy HT plate reader. Gradient of this calibration curve is conversion factor for fluorescence of measured by this plate reader to equivalent ng/ml sodium fluorescein.

For the plate reader: The sodium fluorescein control plate was set up as described previously, and measured with same settings as for characterisation samples. A linear relationship was confirmed (figure 4) showing that it is an accurate way representing absolute fluorescence in terms of equivalent ng/ml of sodium fluorescein, at least for the range covered in this calibration curve. The data was processed as follows: 1) Raw fluorescence data was obtained from plate reader, 2) Subtraction of arithmetic mean average of PBS-only wells, 3) Arithmetic mean average fluorescence calculated for each concentration of fluorescein, 4) Mean fluorescence for the concentration of fluorescein for each concentration measured (FL per ng/ml), 5) Arithmetic mean average of FL per ng/ml gives conversion factor for fluorescence of any sample on the plate reader to equivalent ng/ml of sodium fluorescein. N.B. in the calculation of the conversion factor, the 500 ng and 5 ng points were excluded due to pipetting error accumulation for the 5 ng point being small volumes from the serial dilution, and pipetting error for the 500 ng point which was not the starting well for the serial dilution. Data table can be seen in the appendix.

What quantities were measured?

Plate reader

Green fluorescence (485 ex/528 em), absorbance at 600 nm

Flow cytometer

Green fluorescence (optical filter 530/30, FL1 laser wavelength 488 nm). Also side-scatter (SSC) and forward scatter (FSC) of each cell was recorded as it passed through the laser

How much time did it take to acquire each set of measurements?

Preparation of plates for each machine took between 15 and 30 minutes.

Plate reader

the read for absorbance is under 60 s for the whole plate, and after a bulb-warming period of 180 s the read for fluorescence is also under 90 s

Flow cytometer

Priming the machine can take approximately 10 minutes. Each sample takes 45 seconds, with a wash of ~10 s between wells. For a whole plate of 43 samples in total including controls and beads, this takes approximately 43 minutes. The flow cytometer also records side scatter (SSC) which can be indicative of complexity, and forward scatter (FSC) which for this experiment is irrelevant.

How much does it cost to acquire a set of measurements?

The cost of acquiring these measurements was very low because small aliquots were used of reagents already purchased for other purposes. Similarly the machines used are expensive to purchase and maintain but are used primarily by other users and these experiments were conducted during gaps in this other research. Cost to consider however could include:

  • Reagents: Beads (Sphero™ AccuCount Fluorescent Particles 5.1 uM, ~1E6 particles/ml) (3x 10 ul), Sodium fluorescein (3 x 1 ul), PBS (99 ul x 3), 1x COSTAR plate, 1x black plate Greiner, flow cytometer reagents (system buffer, washing reagents etc).
  • Electricity and maintenance of the cytometer and AMS
  • Time of staff to train in using or to maintain the machines

What are the practical limits on the number or rate of measurements taken with this instrument and protocol?

With the plate reader, one 96-well plate could be measured approximately every 2-3 minutes however accounting for plate preparation time this would be approximately 15 minutes. For the flow cytometer there could be one read approximately every hour for one 96-well plate.

Section III: Measured Quantities


What are the units of the measurement?

The absorbance and fluorescence have arbitrary units. However, after controlling for background fluorescence (plate reader and flow cytometer) and normalising for absorbance (plate reader only), the data is compared to either standard beads (flow cytometer) or sodium fluorescein (plate reader) to obtain fluorescence in terms of these standards. The units from the plate reader will be equivalent ng/ml fluorescein per OD 600 unit, and for the flow cytometer will be fluorescence as a percentage of the standard bead mean fluorescence. This theoretically could be converted into molecules of GFP with a further calibration. OD 600 units can be converted to absolute cell number by calibrating using a flow cytometer, or by creating a standard curve based on counting colonies on plates of dilutions of different OD 600 values.

What is the equivalent unit expressed as a combination of the seven SI base units?

For the plate reader, absolute fluorescence could be expressed as equivalent moles of sodium fluorescein using the MW of 376.27. For the flow cytometer, absolute fluorescence is relative to the standard beads, so an SI unit is not applicable.


What is the range of possible measured values for this quantity, using your instrument as configured for these measurements?

For fluorescence on both the plate reader and the flow cytometer, the sensitivity can be adjusted depending on the sample. For this experiment test wells were done to confirm that for the sensitivity selected, this would include the whole range of the data. For absorbance on the plate reader, readings are accurate between 0 and 1

What are the significant figures for these measurement?

The plate reader measures absorbance to 3 s.f. For fluorescence this reaches 5 s.f. For the flow cytometer the raw data file records 7 s.f., but we used 3-4.

Is the precision the same across the entire range? If not, how does it differ?

The precision is the same across the range for the flow cytometer and plate reader as proven by previous lab experiments where different sensitivity settings were tested.

How did you determine these answers?

For the flow cytometer, this was determined previously in the lab using standard beads of different known fluorescences. For the plate reader, previous tests had compared absorbance readings to a spectrophotometer.


When was the instrument last calibrated?

For the flow cytometer, calibration occurs with every run by using the beads. The plate reader was calibrated approximately 6 months ago.

How was the instrument calibrated?

The beads are run with every flow cytometer plate reading at the end of the experiment. As the beads are commercially available they can serve as a standard for comparing inter-lab variation. For the plate reader, absorbance at 600 nm measurements are compared to a JENWAY Genova spectrophotometer to obtain a standard curve for converting from plate-reader absorbance to OD 600 with a path-length of 1cm.

Section IV: Measurements

Plate Reader: Direct Measurements

Absorbance at 600 nm was measured three times for each of the three cultures, alongside controls for media and contamination background (table 1).

1 2 3 4 5 6 7 8 9 10 11 12
A 0.645 0.592 0.646 0.588 0.496 0.543 0.621 0.581 0.585 0.583 0.608 0.616
B 0.643 0.612 0.644 0.552 0.508 0.528 0.622 0.605 0.623 0.583 0.612 0.593
C 0.633 0.604 0.637 0.529 0.5 0.547 0.599 0.622 0.646 0.629 0.634 0.606
D 0.174 0.164 0.229 0.233 0.2 0.236
Table 1: Absorbance at 600 nm for 3 devices with different promoters coding for GFP as read in a BioTEX Synergy HT plate reader. Units are arbitrary.
  • ‘A’ wells - Existing J23101 device. Three cultures (horizontal) samples three times (vertical)
  • ‘B’ wells - Cloned J23101 device. Three cultures (horizontal) samples three times (vertical)
  • ‘C’ wells - Cloned J23115 device. Three cultures (horizontal) samples three times (vertical)
  • ‘N’ wells - negative control device (no promoter). Three cultures (horizontal) samples three times (vertical)
  • ‘M’ wells - Media plus antibiotic (chloramphenicol only). Sampled three times.
  • ‘O’ wells - Non-inoculated culture (chloramphenicol). Sampled three times

Fluorescence was also measured (table 2), using settings as described previously. The negative cultures account for 7.75% total fluorescence for ‘A’ wells (blue), 1.15% total fluorescence for ‘B’ wells (green) and 23.2% total fluorescence for ‘C’ wells.

1 2 3 4 5 6 7 8 9 10 11 12
A 6102 5914 7191 44832 36996 38835 3182 2940 3023 1625 1519 1436
B 7359 6652 7885 45460 37231 39344 3236 3005 3156 1591 1712 1595
C 7221 6751 7661 43861 36512 39185 3052 3133 3178 1719 1700 1576
D 911 1015 1224 1356 1147 1292 50872 78149
Table 2: Fluorescence (485 ex/528 em) for 3 devices with different promoters coding for GFP as read in a BioTEX Synergy HT plate reader. Units are arbitrary.
  • ‘A’ wells - Existing J23101 device. Three cultures (horizontal) samples three times (vertical)
  • ‘B’ wells - Cloned J23101 device. Three cultures (horizontal) samples three times (vertical)
  • ‘C’ wells - Cloned J23115 device. Three cultures (horizontal) samples three times (vertical)
  • ‘N’ wells - negative control device (no promoter). Three cultures (horizontal) samples three times (vertical)
  • ‘M’ wells - Media plus antibiotic (chloramphenicol only). Sampled three times.
  • ‘O’ wells - Non-inoculated culture (chloramphenicol). Sampled three times
  • Sodium fluorescein control wells. - Note that these wells were not used in analysis of the data in the end, please see the described sodium fluorescein control plate instead for more details.

Plate Reader: Derived Measurements

Data was processed as follows:

  1. Background absorbance was removed by subtracting the average of M and O cells.
  2. These absorbances were multiplied by the conversion factor (8.67) to convert to OD 600 equivalent on a spectrophotometer with a path length of 1 cm. The conversion factor originates from calibration of the plate reader to a spectrophotometer as described previously
  3. Background fluorescence was controlled for by subtracting the average of M and O cells.
  4. Background controlled fluorescence was then divided by background controlled OD 600 to obtain fluorescence per OD unit (FL/OD)
  5. The average FL/OD of the negative control cultures with no promoter was subtracted from the FL/OD of the other constructs to remove background cell fluorescence. The resulting value is fluorescence per OD 600 (arbitrary units).
  6. Dividing these values by the conversion factor as determined from the sodium fluorescein control plate (126.02) gives the absolute fluorescence as the equivalent ng/ml sodium fluorescein per OD unit (table 3, 4). This absolute value should be comparable across different machines which are calibrated in the same way.
A1 A2 A3 B1 B2 B3 C1 C2 C3
I 9.28 10.25 11.52 103.61 112.08 101.3 3.44 3.32 3.48
II 11.96 11.36 13.03 116.16 108.3 107.51 3.55 3.21 3.36
III 11.97 11.84 12.78 119.98 109.03 101.04 3.38 3.32 3.18
Table 3: Absolute fluorescence in terms of ng/ml fluorescein per OD unit of three different GFP encoding constructs with different promoters and plasmid copy number
Arithmetic mean 11.55 108.78 3.36
S.D. 1.17 6.47 0.12
C.V 10.15 5.95 3.56
Ratio 3.44 32.39 1.00
Table 4: Comparison of absolute fluorescence in terms of equivalent ng/ml sodium fluorescein for three different constructs expressing GFP. S.D-standard deviation. C.V-coefficient of variation. Ratio - cultures normalised to C.

Flow Cytometer: Direct Measurements

The negative cultures account for 1.34% total fluorescence for ‘A’ wells (blue), 0.14% total fluorescence for ‘B’ wells (green) and 6.01% total fluorescence for ‘C’ wells. As expected fluorescence is highest for the stronger J23101 promoter (A and B) and for the higher copy number plasmid (B).

A1 A2 A3 B1 B2 B3 C1 C2 C3 N1 N2 N3
I N/A 79.71 70.09 642.79 721.65 670.06 15.1 16.77 16.56 1.02 1.01 1.02
II 73.92 78.99 71.38 676.39 784.49 735.92 15.57 17.82 17.02 1.02 1.02 1.02
III 75.21 79.46 72.19 682.32 703.4 743.01 16.1 18.72 17.75 1.02 1.03 1.01
Table 5: Fluorescence for each well measured using a flow cytometer and determined using geometric mean (A,B,C) or arithmetic mean (N) for gated events in each well. Excluded sample A1-I because voltage was changed whilst sample was running so is not valid. Units are arbitrary.

Flow Cytometer: Derived Measurements

  1. Background fluorescence was removed by subtracting the negative control cultures.
  2. These values were divided by the fluorescence of the beads (1391.51, figure 3C) and multiplying by 100 to obtain the absolute fluorescence as percentage of the standard beads (table 6, 7). This should be comparable across machines using the same standard beads.
A1 A2 A3 B1 B2 B3 C1 C2 C3
I N/A 5.66 4.96 46.12 51.79 48.08 1.01 1.13 1.12
II 5.24 5.60 5.06 48.54 56.30 52.81 1.05 1.21 1.15
III 5.33 5.64 5.11 48.96 50.48 53.32 1.08 1.27 1.20
Table 6: Absolute fluorescence of three GFP devices with different promoters or plasmid copy number, in terms of percentage compared to standard beads measured with a flow cytometer
Mean 5.33 50.71 1.14
S.D. 0.24 2.74 0.08
C.V. 4.51 5.4 6.83
Ratio 4.68 44.5 1
Table 7: Comparison of absolute fluorescence of three GFP devices as measured using a flow cytometer using standard beads as absolute reference control.

Average absolute fluorescence was lowest for construct C (J23115 promoter). The parts registry page (Anderson et al., 2006) shows that J23101 is 4.7 times stronger than J23115. For the supplied J23101 MeasKit our data shows a similar difference of 3.44 times stronger for J23101 than J23115 for the plate reader and agrees with this when the plate reader is used and 4.68 times stronger when the flow cytometer is used (figure 5). There are differences in absolute fluorescence for the two same J23101 constructs A and B because B is in the higher copy number plasmid pSB1C3.

Comparison of Relative Fluorescence for Plate Reader and Flow Cytometer

Figure 5: Ratio of mean GFP fluorescence for each construct compared to the weakest construct for plate reader and flow cytometry methods.

The flow cytometry method gives higher relative fluorescence compared to the J23115 construct than the plate reader method (figure 5).

Extra Credit Assignment 1: Measurement of Cell-to-Cell Variation for the Three Devices

The method of characterising the parts on the flow cytometer measures the fluorescence of each cell individually. In addition to characterising the absolute fluorescence of the different constructs as above, the data can be used to explore the variance of fluorescence between individual cells, since there is a fluorescence reading for each individual cell. This was done by calculating the geometric mean of the whole population for each construct, and the geometric standard deviation (table 8, figure 6). The geometric standard deviation is a good indication of variance in itself because it is a dimensionless multiplicative factor which in essence describes the percentage error from the mean (table 8).

Figure 6. Comparing differences in geometric standard deviation of fluorescence of cells measured on a flow cytometer for three devices expressing GFP.

From our data it can be seen that the variance for the ‘B’ constructs (Cloned J23101 in pSB1C3) is much larger than the other two constructs. However, this could be due to the cells experiencing burden due to the high copy number and high strength promoter, as evidenced by their lower cell number per well (N, table 8). For example the ‘A’ and ‘C’ constructs have at least twice the number of cell events than ‘B’ (table 8). Hence looking at differences in variation may not be relevant as the sample sizes are so different and a direct comparison cannot be made.

Solutions to combat this could be to try and obtain similar OD 600 across all constructs by halting growth of the faster cells by incubating on ice and waiting for the burdened cells to reach the same OD 600. However, this still may not be representative as the lower growth rate can have complex, non-linear and often counter-intuitive relationships with gene expression parameters (Klumpp et al, 2009). It could be possible to try and control growth rate, however this would not be possible except for modifying conditions such as temperature, which would then be an extraneous variable compared to the rest of the constructs to be characterised. A better solution would be to have careful consideration of promoter and backbone copy number, for example a weaker promoter when copy number is high, to ensure burden is not experienced by the cells.

GeoMean GeoStDev N Relative N
A 74.55 1.68 252273 2.12641
B 711.41 2.03 118638 1
C 16.76 1.75 326587 2.752803
Table 8: Comparing cell-to-cell variation of three genetic devices coding for GFP with different promoters and backbones. A- Existing J23101 device, B - Cloned J23101 device, C - Cloned J23115 device.


  1. Anderson, C., Berkley iGEM Team., (2006). Anderson Promoter Collection. Registry of Standard Biological Parts. [accessed 19/09/2014]
  2. Klumpp, S., Zhang, Z., Hwa, T., (2009). Growth rate-dependent global effects on gene expression in bacteria. Cell 139(7):1366-75. doi: 10.1016/j.cell.2009.12.001.


Raw data for fluorescein control plate

Raw data for fluorescein control plate (with background DPBS diluent subtracted) to determine conversion factor of fluorescence to equivalent ng/ul of sodium fluorescein. Conversion factor was 126.02, the average of FL/(ng/ml). As noted previously, the 500 ng and 5 ng readings were excluded from the calculation as were deemed too inaccurate due to pipetting error.

ng/ml Replicate I Replicate II Replicate III Mean S.D. C.V. FL/(ng/ml)
500 39779 43169 58234 47061 9823.72 20.87 94.12
375 39223 42055 54108 45129 7904.2 17.51 120.34
250 25511 29221 36966 30566 5844.74 19.12 122.26
125 13805 15193 18957 15985 2665.75 16.68 127.88
50 5670 6042 7878 6530 1182.13 18.1 130.6
25 2812 2967 4012 3264 652.69 20 130.55
10 1022 1180 1533 1245 261.63 21.01 124.5
5 554 808 776 713 138.34 19.41 142.53
0 2 -2 0 0 2 N/A N/A
Table 9: Fluorescence of different concentrations of sodium fluorescein (ng/ml) as measured in a BioTEX Synergy HT plate reader with blank DPBS control subtracted. Units are arbitrary. FL. - fluorescence, S.D.- standard deviation, C.V. coefficient of variation (all arithmetic)

Example flow cytometry graphs for ‘A’ (Existing J23101 construct) well F1.

Figure 7: Dot plot of side scatter (SSC) vs forward scatter (FSC) showing population (red, blue) SSC gate for excluding cell debris.
Figure 8: Histogram of SSC before gating
Figure 9: Histogram of SSC after gating
Figure 10: Dot plot of SSC vs fluorescence (FL1) showing gating to exclude cell debris (blue)
Figure 11: Histogram of gated FL-1 showing log-normal distribution

Example flow cytometry graphs for ‘B’ (Cloned J23101 construct) well F4.

Figure 12: Dot plot of side scatter (SSC) vs forward scatter (FSC) showing population (red, blue) SSC gate for excluding cell debris.
Figure 13: Histogram of SSC before gating
Figure 14: Histogram of SSC after gating
Figure 15: Dot plot of SSC vs fluorescence (FL1) showing gating to exclude cell debris (blue)
Figure 16: Histogram of gated FL-1 showing log-normal distribution

Example flow cytometry graphs for ‘C’ (Cloned J23115 containing the mismatches of K823012 construct) well G9.

Figure 17: Dot plot of side scatter (SSC) vs forward scatter (FSC) showing population (red, blue) SSC gate for excluding cell debris.
Figure 18: Histogram of SSC before gating
Figure 19: Histogram of SSC after gating
Figure 20: Dot plot of SSC vs fluorescence (FL1) showing gating to exclude cell debris (blue)
Figure 21: Histogram of gated FL-1 showing log-normal distribution

Example flow cytometry graphs for ‘N’ (construct with no promoter i.e. E0240 by itself) well F10.

Figure 22: Dot plot of side scatter (SSC) vs forward scatter (FSC) showing population (red) SSC gate for excluding cell debris.
Figure 23: Histogram of SSC before gating
Figure 24: Histogram of SSC after gating
Figure 25: Dot plot of SSC vs fluorescence (FL1). The population of non-fluorescing cells and debris is indistinguishable from debris so no gating is used.
Figure 26: Histogram of FL-1