Although DSBs are considered one of the most critical types of DNA damage, studies show that DSBs comprise only 20–30% of the total yield of multiply damaged sites (MDS).  The repair of MDS other than the DSB is expected to be error prone and contribute substantially to radiation-induced mutagenesis (Ward et al. 1985, Ward 1994, Ward 1995). The underlying hypothesis is that MDS repair is error prone because repair enzymes may not always have an intact template to guide the replacement of damaged nucleotides (Ward 1988, Goodhead 1994).  Evidence for the conversion of MDS into potentially lethal DSBs is also accumulating (Yang et al. 2004).  DSBs may be formed through the aborted excision repair of some types of MDS.

One of the new features in VC Version 2.0 is the ability to simulate for the formation, repair and misrepair of 2 kinds of singly damaged DNA site (single base damages and a single strand break) and 100 kinds of MDS, including one (LPL and RMR model) or two kinds of DSB, multiply based damaged sites and complex single strand breaks (SSB with additional base damage or strand breaks on the same DNA strand).

The attempt to simulate biological processes in greater detail inevitably introduces additional parameters into the modeling process. To simulate the formation, repair and misrepair of 100 different kinds of DNA damage introduces a minimum of 300 additional parameters into the modeling process. Each type of damage is modeled using a differential equation of the form

dL_i(t)/dt = S_i·ADR(t) - l_iL(t)

Here, S_i is the initial yield of the ith kind of damage per Gy per cell and l_i characterizes the repair rate for the ith kind of damage, and ADR(t) is the absorbed dose rate at time t.  A third parameter, the probability of correct repair a_i, must be introduced to determine the fraction of the initial damage converted to a lethal or non-lethal mutation.  Three adjustable parameters per damage type times one hundred kinds of damage equals three hundred parameters.

A model with three hundred (or more) purely adjustable parameters is impractical, if the parameters can only be estimated by fitting the model to measured data.  This issue is sometimes referred to as the "Give me enough parameters, and I can fit an elephant!" problem.  To help overcome these parameter estimation issues, we have developed and benchmarked two other computer programs to predict S_i and a_i from more fundamental physical and biological considerations.  The Monte Carlo Damage Simulation (MCDS) algorithm (Semenenko and Stewart 2004) provides estimates of the initial yield of DNA damage per Gy per cell (S_i parameter), and the Monte Carlo Excision Repair (MCER) model (Semenenko et al. submitted) provides pathway-specific estimates of the probability a damage configuration is repaired correctly (a_i parameter), as well as other useful information about repair outcomes.

1. Paste the contents of RMR sample file #2 into a new ASCII input file called rmr1.inp.
2. Verify that the PHI parameter (probability a point mutation is lethal) is set to 0.001.
3. Run the simulation from the command line: type vc rmr1.inp

• Towards the top of the output file, is this line of text 'MCER: c:\vc\bxs\electron.mcer' present? If not, the VC environment variable may not be set correctly. For helping setting up the VC environment variable, see the System Configuration section of the manual. Alternatively, please verify that the electron.mcer file is present in the .\vc\bxs folder.
• Correct any setup/installation problems before continuing with the exercise.

• Open the rmr1.out file.  What fraction of the LQ parameter a is due to singly and multiply damaged DNA sites other than the DSB (Hint: See the 'Other Lesions' portions of the LQ table)? Use information from DFAR TABLE 1 and DFAR TABLE 2 to verify the Other Lesion portion of a.
• Set the value of PHI to 0.01 and re-run the simulation.  As the value of PHI, increases does the other lesion portion of a increase, decrease or stay the same?
• How many singly and multiply damaged DNA sites (other than the DSB) are created per Gy per cell?   See DFAR TABLE 1.
• How many non-lethal single strand breaks (SSBs) are formed per Gy per cell?  Use information from DFAR TABLE 1 and DFAR TABLE 2 to estimate the yield of non-lethal SSB.

1. Create a copy of the rmr1.inp file.  Rename the file rmr2a.inp. 
2. Set the value of PHI to 0.001
3. Run the rmr2a.inp simulation.  At the command prompt type: vc rmr2a.inp electron.mcer (low LET radiation)
4. Create copy of the rmr2a.inp file and rename it rmr2b.inp
5. Run the rmr2b.inp simulation.  At the command prompt
   type: vc rmr2b.inp alpha2MeV.mcer (high LET radiation)

• Is the surviving fraction for the a particle (rmr2b.out) higher, lower or the same as the surviving fraction reported for the low-LET electron (rmr2a.out)?
• Compare estimate of a from the rmr2a.out and rmr2b.out files.  Is value of a for the high LET radiation (rmr2b.out) higher, lower, or the same as the a for the low LET radiation (rmr2a.out).  If a increases or decreases with LET, explain the reason for the change, i.e., is the change due to 'other lesions', DSB or both?
• Compare b reported in the rmr2a.out and rmr2b.out files.  Is value of b for the high LET radiation (rmr2b.out) higher, lower, or the same as the a for the low LET radiation (rmr2a.out)? 
• Did the SSB yield per Gy per cell increase, decrease or stay the same as particle LET increased?