VHEE FLASH sparing effect measured at CLEAR, CERN with DNA damage of pBR322 plasmid as a biological endpoint

Ultra-high dose rate (UHDR) irradiation has been shown to have a sparing effect on healthy tissue, an effect known as ‘FLASH’. This effect has been studied across several radiation modalities, including photons, protons and clinical energy electrons, however, very little data is available for the effect of FLASH with Very High Energy Electrons (VHEE). pBR322 plasmid DNA was used as a biological model to measure DNA damage in response to Very High Energy Electron (VHEE) irradiation at conventional (0.08 Gy/s), intermediate (96 Gy/s) and ultra-high dose rates (UHDR, (2 × 109 Gy/s) at the CERN Linear Electron Accelerator (CLEAR) user facility. UHDRs were used to determine if the biological FLASH effect could be measured in the plasmid model, within a hydroxyl scavenging environment. Two different concentrations of the hydroxyl radical scavenger Tris were used in the plasmid environment to alter the proportions of indirect damage, and to replicate a cellular scavenging capacity. Indirect damage refers to the interaction of ionising radiation with molecules and species to generate reactive species which can then attack DNA. UHDR irradiated plasmid was shown to have significantly reduced amounts of damage in comparison to conventionally irradiated, where single strand breaks (SSBs) was used as the biological endpoint. This was the case for both hydroxyl scavenging capacities. A reduced electron energy within the VHEE range was also determined to increase the DNA damage to pBR322 plasmid. Results indicate that the pBR322 plasmid model can be successfully used to explore and test the effect of UHDR regimes on DNA damage. This is the first study to report FLASH sparing with VHEE, with induced damage to pBR322 plasmid DNA as the biological endpoint. UHDR irradiated plasmid had reduced amounts of DNA single-strand breaks (SSBs) in comparison with conventional dose rates. The magnitude of the FLASH sparing was a 27% reduction in SSB frequency in a 10 mM Tris environment and a 16% reduction in a 100 mM Tris environment.

upgraded to VHEE beam energies 31,[33][34][35] .150 and 201 MeV electrons were achievable and repeatable using the CLEAR facility.It was important to explore energies within the range of VHEE to add to the very limited body of research with electrons over ~ 35 MeV.In regards to the VHEE energies used in this study, the linear energy transfer (LET) is ~ 0.6-0.7 keV/μm for electrons in the 150-201 MeV range 36 .This is slightly higher than clinically used energies (~ 0.2 keV/μm for 10 MeV electrons).Results from other studies indicate that VHEEs have a relative biological effectiveness (RBE) of approximately 1, or higher, with comparison to a photon reference.Small et al. 29 used a plasmid model to define VHEE RBE as ~ 1.1-1.2 for electrons in the 100-200 MeV range.Simulation studies have calculated VHEE RBE as 0.99-1.03 37,38.
Irradiations for this study took place at the CERN Linear Electron Accelerator for Research (CLEAR).The CLEAR electron linac offers users beam energies of 60-220 MeV 39 and flexible beam parameters that result in a unique ability to alter dose rate regimes from conventional to several orders of magnitude above the FLASH 'threshold' of ~ 100 Gy/s 1 .Recent developments to the CLEAR infrastructure, including a robot to handle samples, meant that the plasmids were irradiated with high efficiency and precision 39 .The ability to irradiate up to 32 plasmid samples per batch meant that two concentrations of hydroxyl scavenger Tris could be tested, over a dose range of ~ 15-200 Gy.The unique parameters at CLEAR allowed us to investigate the response of plasmid DNA to different dose rate regimes, at two VHEE energies.

Materials and methods pBR322 plasmid samples
All samples were prepared to total volume of 8 μl within a 0.5 ml Eppendorf tube (Eppendorf, 0030121023), containing 100 μg/ml pBR322 plasmid DNA (New England Biosciences, N3033) and either 10 or 100 mM Tris hydroxyl scavenger (Thermo Fisher Scientific, AN9855G).Distilled water was used to dilute the sample to its final concentration of pBR322 and Tris.The stock plasmid from New England Biosciences contains a residual concentration of Tris-HCl which was included in the calculation of the final concentration of Tris hydroxyl scavenger.The stock plasmid also contains residual EDTA which acts to prevent the degradation of DNA by the deactivation of DNase 40 .10 and 100 mM concentrations of Tris represent hydroxyl scavenging capacities that are approximately 4% and 40% of a cellular hydroxyl scavenging capacity respectively.This is based on the hydroxyl scavenging capacity of pure Tris (1.1 × 10 9 /M s), which can be used to calculate hydroxyl scavenging capacities of 1.1 × 10 7 /s and 1.1 × 10 8 /s for 10 mM and 100 mM Tris environments respectively [41][42][43] .Comparisons can then be made with cellular hydroxyl scavenging capacity, taken as 3.0 × 10 8 /s 43 .

Irradiation of samples at CLEAR, CERN
All samples were irradiated at the CLEAR facility.Samples were processed using gel electrophoresis at the Oglesby Cancer Research Building (OCRB) located in Manchester, UK because facilities to process the plasmid DNA with gel electrophoresis were not available at CERN.Plasmid samples were transported to and from the OCRB on ice and were stored at − 20 °C at all times other than during sample preparation, transport and irradiation.All samples were irradiated with the plasmid solution at equilibrium at room temperature and air conditions (20% oxygen).
Samples were presented to the beam for irradiation using the C-Robot, a device that was used to reduce the turnover time of sample irradiation and number of accesses to the accelerator hall.The C-Robot was controlled by the operator, allowing individual samples to be 'grabbed' with high precision and held in front of the beam during the irradiation time.Custom sample holders (Fig. 1) were designed and 3D printed by the CLEAR team to fit the 0.5 ml Eppendorf tubes containing the sample volume.
The samples holders consisted of a round slot in which to place the Eppendorf tube, a 'handle' for the C-Robot to grab onto, and two slits for Gafchromic film to be placed in front and behind each sample.EBT-XD Gafchromic film 44 was used for doses up to 60 Gy.MD-V3 Gafchromic film 45 was used for samples irradiated in the 60-200 Gy dose range.Up to 32 samples were irradiated in each batch, minimising the frequency of access to the accelerator hall required.
Samples were moved to varying depths (1-10 cm) within the water tank to alter beam size.Electron range in water is indicated below in Fig. 2 for context, to show the penetration depth of 150 and 201 MeV electrons using a percentage dose depth curve.
The beam in air before entry to the water tank was a symmetrical Gaussian beam, σ = 0.8-1.2mm.The beam at the sample depth consisted of a symmetrical Gaussian beam, σ = 2.8-5.5 mm.In regards to electron fluence, a 5.5 × 5.5 mm σ beam size required ~ 1 nC to achieve 1 Gy.For a 5.0 × 5.0 mm σ beam size, ~ 0.8 nC was required to achieve 1 Gy.Experimental set-up is indicated below in Fig. 3.

Calculation of sample dose and dose rate
Dose was measured from either EBT XD or MDV3 Gafchromic film by analysing within the irradiated sample area.The sample area was located on the film images by using example films which were placed behind a 0.5 ml Eppendorf tube containing a small lead pellet, which was then irradiated.The lead pellet absorbed a significant amount of dose so that a visible spot could be located on the film.This spot was then used to determine where the bottom of the Eppendorf tube, and therefore the sample was located.Once the location of the sample area was isolated, the mean dose was measured from the film placed directly behind and in front of each sample.Average dose rate was calculated for conventional, intermediate and ultra-high dose rate (UHDR) by measuring the dose to sample with Gafchromic film and dividing by the total irradiation time.Average dose rates were calculated as follows: 0.08 Gy/s (conventional), 96 Gy/s (intermediate) and of the order of 2 × 10 9 Gy/s (UHDR).Physical parameters are defined for each dose rate used below in Table 1.www.nature.com/scientificreports/

Separation of plasmid structures with gel electrophoresis
Irradiation of plasmid DNA within the dose range used for this experiment results in damage in the form of single-and double-strand DNA breaks, known as SSBs and DSBs respectively.The native structure of pBR322 plasmid is supercoiled (SC).Plasmids that have been damaged by a SSB or DSB will subsequently change structure.Plasmids with damage in the form of an SSB will display an open circular (OC) structure, whereas a more damaging DSB will result in a linear (L) structure.SC, OC and L structures can be therefore be separated by gel electrophoresis due to their subsequent differences in plasmid mobility.All plasmid samples were separated using gel electrophoresis methods.This was completed at the Oglesby Cancer Research Building (OCRB), UK. 1% w/v agarose gel was prepared with 1× TAE buffer (Thermo Fisher Scientific, B49), agarose powder (Sigma Aldrich, A9539) and 1× SYBR Safe DNA gel stain (Thermo Fisher Scientific, S33102).To process the samples with gel electrophoresis, 83.3 μg/ml pBR322 plasmid DNA and 1X loading dye (Thermo Fisher Scientific, R0611) is loaded into the gel at a total volume of 1 μl.Once set, the gel was placed in an electrophoresis tank and covered in 0.5× TAE buffer (Thermo Fisher Scientific, B49).A constant voltage of 70 V was then applied for 2.5 h.
Gels were imaged on the ChemiDoc MP UV imager (BioRad) using the 'SYBR Safe' setting, with the exposure condition always set to ' Automatic Rapid Exposure' .

Quantification of DNA damage with the McMahon fit
To quantify the relative amount of plasmid DNA in each proportion, the integrated density of each band within the electrophoresis gel was measured using Image J software.The number of induced SSBs and DSBs was then calculated using the McMahon and Currell fitting procedure 46 .The curves were fit to data points using Microsoft Excel Solver (2016) by calculating curves that minimised the least squared fit.Values of β s and β D (SSB and DSB respectively) were calculated by minimised the square of the error for the measured SC, OC and L proportions and the McMahon and Currell fit.
where SC, OC and L are the proportions of supercoiled, open circular and linear plasmid structures respectively.SC 0 , OC 0 and L 0 are the mean proportions of supercoiled, open circular and linear plasmid structures in unirradiated controls for 10 and 100 mM Tris concentrations as applicable.D represents the dose to sample in Gy. ρ represents the ratio between the length of pBR322 plasmid (4361 bp) and the maximum distance between two SSBs resulting in a DSB (taken as 10 bp 29,47 ).β s and β D are the frequencies of SSBs and DSBs per unit dose across the range of doses used to irradiate the plasmid samples.
Samples were irradiated in various conditions during the experiment, with several doses for each condition.In the majority of cases, enough dose points were used so that a McMahon fit could be made and SSB and DSB frequencies could be calculated with a degree of statistical rigour.In the case of the intermediate dose rate irradiations, irradiations were only successfully completed at two doses and therefore the McMahon curve was not utilised to calculate SSB or DSB frequencies for this data.This was because the limited number of doses was not deemed to be as statistically rigorous as other conditions with a larger number of dose points.The intermediate data at two doses is plotted in Fig. 5, however a McMahon fit has not been plotted alongside for this reason.

Statistical analysis
Values of SSB induction for each condition were calculated using McMahon fits as described above.The 95% and 68% confidence intervals (CI) were calculated on each SC and OC McMahon fit using a custom Python code.The weighted 95% CI that considered both SC and OC fits is presented in Table 2 and is plotted as the error values in Figs. 8 and 9.
(1) To determine statistical significance, an Unpaired t-test was completed, where a P-value of < 0.05 was considered significant.For the t-test, the 68% CI on either the SC or OC McMahon curve was taken as the standard deviation, whichever was the more significant error.N was taken as the number of irradiated samples for each curve where N = 8-12.

Results
Plasmid samples were irradiated at various doses (up to ~ 200 Gy).All pBR322 plasmid structures were then separated via gel electrophoresis as previously described.Figure 4 shows an example of how plasmid structure changes in response to dose.Over 150 plasmid samples were processed this way, therefore Fig. 4 shows a small proportion of the samples, as an example only.Figure 4 shows three unirradiated samples as well as a selection of samples irradiated with conventional 201 MeV electrons in a 10 mM Tris environment.
It can be seen that for the unirradiated (0 Gy) plasmid samples, the majority of plasmid remains in the native SC structure, where band intensity is visibly strong.Fainter bands can be seen in the OC structure, indicating a smaller percentage of the plasmid is in this structure.No visible band can be seen in the L structure, suggesting no or very few DSBs have occurred in the unirradiated sample.This is verified by measuring the integrated density of each plasmid structure in its unirradiated form.Unirradiated samples within a 10 mM Tris environment were measured to have an average SC proportion of 0.89 ± 0.04 with a corresponding OC proportion of 0.08 ± 0.03, and a smaller L proportion of 0.03 ± 0.03.Errors represent standard deviation.Increasing the concentration of Tris to 100 mM was shown to significantly increase the proportion of plasmid remaining in its native SC structure, protecting the plasmid from base levels of damage during transport and storage.Unirradiated samples maintained within a 100 mM Tris environment had an average SC proportion of 0.98 ± 0.03, an OC proportion of 0.02 ± 0.03 and a corresponding L proportion of 0.001 ± 0.003.
Figure 4 also shows how the quantity of plasmid structures changes when irradiated with doses in a range of 28.5-104.0Gy.Post-irradiation, more intense bands can be observed for the OC structure in comparison to unirradiated plasmid indicating a higher frequency of SSBs for irradiated plasmid.Differences cannot be easily distinguished between the proportions of plasmid in OC and L structures at each dose.It can however be identified with visual inspection that the quantity of plasmid in SC structure decreases with increased dose based on the band intensity.This is quantified for each experimental condition in Figs. 5, 6 and 7.
The proportions of pBR322 plasmid structure (SC, OC and L) for every experimental condition have been plotted in response to dose (D) in Figs. 5, 6 and 7. Figure 5 shows the response of samples irradiated with 201 MeV electrons, in an environment of 10 mM Tris.The structure proportions of pBR322 for conventional, intermediate and ultra-high dose rate (UHDR) have been plotted to compare between dose rates.McMahon fits for conventional and UHDR conditions have also been included.
Figure 6 shows the response of pBR322 plasmid irradiated with 201 MeV electrons, in a 100 mM environment.pBR322 structure proportions for conventional and UHDRs are indicated as well as the corresponding McMahon fits.
Irradiations were completed at two VHEE energies: 150 and 201 MeV.www.nature.com/scientificreports/Data summarised in Table 2 has been plotted in Fig. 8 to compare between the DNA damage induced by conventional and UHDR irradiations.SSB frequency is the biological endpoint used to perform the statistical t-test due to the significantly higher frequency of SSBs than DSBs. Figure 8a and b show the SSB induction to conventional and UHDR irradiations in a 10 mM and a 100 mM Tris environment respectively.Based on the  response of an unpaired t-test, SSB induction is lower in response to UHDR irradiation than conventional irradiation in both conditions of Tris hydroxyl scavenger.
Figure 9 shows the SSB induction in response to irradiations of two different electron energies: 150 MeV and 201 MeV.Both electron energies were irradiated at conventional dose rates within a 10 mM Tris environment.A t-test indicated that 150 MeV electrons had significantly lower SSB frequency in comparison to 201 MeV electrons.
Proportional values have been calculated from the data shown in Table 2 above, as an indication of the DNA damage inflicted by ultra-high dose rate, and for 150 MeV electron irradiations (in comparison to 201 MeV, conventional dose rate electrons).These proportional values are shown in Table 3.

Discussion
The aim of this study was to investigate the biological effect of dose rate and beam energy in the context of VHEE.The results show that UHDR irradiation significantly reduces SSB induction in comparison to conventional irradiation.This is the case in two hydroxyl scavenging environments, 10 mM and 100 mM Tris.Increased concentration of Tris scavenger resulted in a significant reduction in damage to the plasmid DNA due to the reduction of hydroxyl-mediated indirect DNA damage.The results also indicated that the SSB induction was significantly increased in plasmid samples irradiated with 150 MeV electrons, in comparison to plasmid irradiated at 201 MeV.In a 10 mM Tris environment, pBR322 plasmid irradiated with 201 MeV electrons resulted in a SSB frequency of 3.05 ± 0.08 (conventional) and 2.24 ± 0.14 (UHDR) (/Mbp Gy).Corresponding DSB frequencies of 0.06 ± 0.03 (conventional) and 0.05 ± 0.07 (UHDR) (/Mbp Gy) were also observed.At 10 mM Tris, a set of irradiations were also completed using a 150 MeV electron beam.At 150 MeV, the SSB and DSB frequencies were 3.40 ± 0.17 and 0.11 ± 0.07 (/Mbp Gy) respectively for conventional irradiated plasmid.
In the 100 Mm Tris environment, the frequency of SSBs was reduced significantly.In the 100 mM environment with 201 MeV electron irradiation, the SSB frequency 0.68 ± 0.07 (conventional) and 0.57 ± 0.21 (UHDR) and the DSB frequency was 0.04 ± 0.04 (conventional) and 0.04 ± 0.10.Due to the high error on the DSB frequencies, that data has not been plotted and the endpoint of SSB frequency has been used to compare between dose rates and beam energies.This is the first time that FLASH sparing has been reported using VHEE, with induced damage to pBR322 plasmid DNA as the biological endpoint.The results of this study are in contrast to those published by Small et al. 29 in 2021, where plasmid samples were also irradiated at the CLEAR user facility, at conventional and UHDR.There are some minor differences in the methodology between this study and those published by Small et al., the main difference being that this study used a Tris concentration in a 10-100 mM range whereas Small et al., used a 1 mM Tris concentration throughout.Another key difference is that Small et al., used DSB frequency as the biological endpoint whereas for this study, it was determined that the statistical frequency of DSBs was not large enough to determine significant differences between dose rates.Therefore DSBs were not used to make these comparisons.www.nature.com/scientificreports/On the other hand, the results of this study are supported by two other studies that have observed DNA damage sparing to plasmid in response to 27.5 MeV protons 27 , and 16 MeV electrons 28 .Ohsawa et al., reported that UHDR (40 Gy/s) protons resulted in sparing of pBR322 plasmid damage where SSB frequency was the biological endpoint measured.The plasmids were irradiated in an environment containing 10 mM Tris-HCl.Similar to this study, Ohsawa et al., found that UHDR irradiation reduced DSB frequency however this was not to a statistically significant level.
This could indicate that the biological FLASH effect does not translate to DSBs, suggesting that the most lethal form of damage is still maintained, with a reduction in only sub-lethal damages.Another interpretation is that the significantly lower frequency of DSBs in comparison to SSBs means that it is not statistically viable to determine a difference using DSB frequency as the biological endpoint using published methods.To fully understand the effect of dose rate on DSB induction, larger volumes of plasmid, higher doses or the removal/ reduction of hydroxyl scavengers from the plasmid environment could be used to increase DSB frequency.These alterations to the protocol could be tested in further studies, but these changes could be limited by cost, beam time duration or dosimetry, if using doses out of the range of Gafchromic film.Overall, the study by Ohsawa et al., indicates that the FLASH effect takes place in a plasmid model, and that this is independent of modality, as has also been shown in vivo.
Perstin et al., tested the effect of 16 MeV electrons at conventional and UHDR (46.6 and 93.2 Gy/s) in an environment of 0.48 mM Tris-HCl.It was determined from this study that the FLASH sparing effect was significant at both dose rates, with a more pronounced effect for SSB induction than for DSBs, even for this low concentration of Tris scavenger.
Overall, these collective studies indicate that using the pBR322 plasmid model is an effective way to identify factors that impact DNA damage sparing, providing insight into measuring the biological FLASH effect.This information should be used to inform more complex experiments such as in vitro and in vivo work to elucidate a biological mechanism for FLASH.
In vitro studies have mixed results when using biological endpoints to measure the FLASH effect.A recent review 26 investigated the current breadth of research in this area, reporting that 6 out of 19 studies measuring cell survival showed evidence of the FLASH effect in any condition (in hypoxic or normoxic conditions, or in cancer or normal cells).Despite the varying reports of FLASH in vitro, there is evidence of UHDRs resulting in sparing of DNA damage linked biological effects.UHDRs have been shown to reduce dicentric yields in irradiated human blood 48 .Reduced levels of DNA damage markers including γH2AX and 53BP1 have also been observed in both tumour and normal cells with UHDRs, in comparison to conventional dose rates 5,49 .These studies support the hypothesis that one of the major mechanisms for the FLASH effect is by a reduction in DNA damage, however more mechanistic information is required to fully understand the biological FLASH mechanism from irradiation to in vivo impact.
The FLASH effect has now been observed in plasmids within Tris environments from 0.48 to 100 mM, and in response to protons, electrons and VHEE.Plasmid DNA can and should be used to test a wide range of UHDR regimes to get an indication for the necessary parameters for the FLASH effect.These parameters could then be used to guide in vivo studies.
It should be noted however that a measuring SSB frequency alone does not paint a full picture of the biological FLASH effect.The quantity of DNA damage is important, however it must be used as a first step to facilitate further studies on how DNA damage affects FLASH endpoints that occur at later time points, including DNA repair, cell cycle and cell survival.

Figure 1 .
Figure 1.(a) Image of the 3D printed sample holders used during the irradiations.0.5 ml Eppendorf tubes were placed through the entry at the top of the sample holder to secure the tube.Gafchromic films (40 × 35 mm) were then slotted in front and behind each sample to validate dosimetry measurements.(b) Scanned Gafchromic film image with reference to the Eppendorf tube and beam spot (represented by the darkening on the film).

Figure 2 . 8 Figure 3 .
Figure 2. The percentage on-axis dose deposited in respect to depth within a 30 × 30 × 30 cm water phantom for 150 and 201 MeV electrons.Simulation was completed in TOPAS using 10 6 electrons for each curve, with a Gaussian beam source of σ = 3 mm.Dose was measured across 150 Z axis bins (2 mm × 2 mm × 2 mm each).

Figure 7
represents the effect of these two energies on the pBR322 plasmid structure by comparing conventional irradiated samples, in a 10 mM Tris environment.A summary of SSB (β S in Eqs.1-3) and DSB (β D ) frequencies (/Mbp Gy) for each experimental condition is summarised below in Table2alongside the confidence in corresponding McMahon fits.

Figure 4 .
Figure 4. Image of agarose gel of pBR322 plasmid samples irradiated with conventional (CONV) 201 MeV electrons, in a 10 mM environment.Plasmid DNA is separated out into three structures: Open circular (OC)top band, red.Linear (L), middle band, green.Supercoiled (SC)-bottom band, blue.The image represents the original image as taken with no processing.Dose (Gy) is indicated for each band on the gel by the labelling on the bottom of the figure.Unirradiated samples are indicated by the three lanes labelled 0.0 Gy.Each of the displayed samples were derived from the same experiment and gels were processed in parallel.Gels were imaged on the ChemiDoc MP UV imager (BioRad) using the 'SYBR Safe' setting, with the exposure condition always set to ' Automatic Rapid Exposure' .This figure is a cropped section of a gel-the full gel image can be seen in the supplementary section.

Figure 5 .
Figure 5. Relative proportions of SC (red), OC (blue) and L plasmid (green) relative to dose (D) in Gy within a 10 mM Tris environment.Irradiations were completed with 201 MeV electrons.The effect of dose rate on DNA damage is indicated here, with conventional, intermediate and ultra-high dose rate (UHDR) representing dose rates of 0.08, 96 and in the order of 2 × 10 9 Gy/s respectively.The solid and dotted lines represent the McMahon fits to the conventional and UHDR data sets respectively.No fit has been added for the intermediate dose rate because the data set only included two doses which was not deemed enough points to make a reliable fit.Points represent the mean dose and plasmid proportions of 2 or 3 irradiation repeats.Error bars represent standard deviation (SD) in both X and Y.

Figure 6 .
Figure 6.Relative proportions of SC (red), OC (blue) and L plasmid (green) relative to dose (D) in Gy within a 100 mM Tris environment.Irradiations were completed with 201 MeV electrons.The effect of dose rate on DNA damage is indicated here, with conventional and ultra-high dose rate (UHDR) representing dose rates of 0.08 and in the order of 2 × 10 9 Gy/s respectively.The solid and dotted lines represent the McMahon fits to the conventional and UHDR data sets respectively.Points represent the mean dose and plasmid proportions of 2 or 3 irradiation repeats.Error bars represent standard deviation (SD) in both X and Y.

Figure 7 .
Figure 7. Relative proportions of SC (red), OC (blue) and L plasmid (green) relative to dose (D) in Gy within a 10 mM Tris environment.Irradiations were completed with either 150 or 201 MeV electrons at conventional dose rates.The effect of beam energy on DNA damage is indicated here, with solid and dotted lines representing the McMahon fits to the 201 MeV and 150 MeV irradiated sets respectively.Points represent the mean dose and plasmid proportions of 2 or 3 irradiation repeats.Error bars represent standard deviation (SD) in both X and Y.

Figure 8 .
Figure 8.The effect of dose rate on SSB induction is shown here in both a 10 mM Tris environment (a) and 100 mM Tris environment (b).Conventional and ultra-high dose rate (UHDR) representing dose rates of 0.08 and in the order of 2 × 10 9 Gy/s respectively.Error bars represent the weighted 95% confidence intervals of the McMahon fits to the corresponding SC and OC curves, as shown in Table 2. **** = statistical difference where P ≤ 0.0001, * = statistical difference where 0.01 ≤ P ≤ 0.05.

Figure 9 .
Figure 9.The effect of electron energy on SSB induction is shown here in a 10 mM Tris environment.Irradiations took place at a conventional dose rate with a beam energy of either 150 or 201 MeV.Error bars represent the weighted 95% confidence intervals of the McMahon fits to the corresponding SC and OC curves, as shown in Table 2. **** = statistical difference where P ≤ 0.0001.

Table 1 .
Physical parameters of the CLEAR beamline for each dose rate used to irradiate plasmid samples.

Table 3 .
Proportional values of SSB frequency calculated from values of SSB (β S ) induction as provided in

Table 2 .
29lues were calculated using SSB INTEREST /SSB REFERENCE, where SSB INTEREST is the frequency of SSBs (/Mbp Gy) induced by the radiation of interest (UHDR or 150 MeV conventional)29.SSB REF describes the frequency of SSBs (/Mbp Gy) induced by the reference radiation type (201 MeV, conventional dose rate irradiation).Only values of the same Tris concentration were used to calculate proportional values, e.g.where values was calculated for UHDR in a 10 mM environment, the reference used would be conventional irradiated samples in a 10 mM environment.