Geant4 - clustering
Authors: Y. Perrot (a), H. Payno (b)
(a) ysperrot@gmail.com
(b) henri.payno@gmail.com
Laboratoire de Physique Corpusculaire de Clermont-Ferrand, CNRS/IN2P3 - Clermont University, France
This example is provided by the Geant4-DNA collaboration.
These processes and models are further described at:
geant4-dna.org">http://geant4-dna.org
Any report or published results obtained using the Geant4-DNA software shall
cite the following Geant4-DNA collaboration publications:
Phys. Med. 31 (2015) 861-874
Med. Phys. 37 (2010) 4692-4708
---> 1. Introduction
The clustering example simulates protons tracks in liquid water using Geant4-DNA processes and models.
Energy deposit are clustered with a dedicated clustering algorithm to assess strand breaks.
The default parameters of the clustering algorithm have been tuned to reproduce data published by
Francis et al. 2011 Comput. Meth. Programs. Biomed. 2011 101(3)
Any report or published results obtained using the Geant4-DNA software shall
cite the following Geant4-DNA collaboration publication:
Med. Phys. 37 (2010) 4692-4708
----> 2. Set-up
It is similar to the geometry set-up proposed in Francis et al. 2011 Comput. Meth. Programs. Biomed. 2011 101(3).
It consists in a World volume containing a Target box made of liquid water of 1µm x 1µm x 0.5 µm.
Energy deposits in the Target are registered (see SteppingAction.cc) and the clustering algorithm is run at the end of each event
(see EventAction.cc)
----> 3. How to run the example
To get help, run:
> ./clustering -h
In interactive mode, run:
> ./clustering -gui
In batch mode , run:
> ./clustering [-mac run.in] [-mt numberofThreads]
Two macros are available:
run.in: shoots 1000 protons of 500 keV
runOneEvent.in: shoots one proton of 500 keV
All UI clustering commands in these macros are described below in section 5.
----> 4. Simulation output
The output results consists in a clusters_output.root file, containing for each event:
- the number of single strand break
- the number of complex single strand break
- the number of double strand break
- the cluster size distribution
- the absorbed dose in the Target
----> 5. More information
Specific classes are available in this example:
* ClusteringAlgo: contains the core clustering algorithm
* ClusteringAlgoMessenger: defines all UI commands to tune
the clustering algorithm
/clustering/algo/setMinPts:
Minimal number of points to create a cluster
/clustering/algo/setSelectionProb:
Probability to select potential damage according to the geometry
/clustering/algo/setEps:
Maximal distance between points to create a cluster
/clustering/algo/setEmin:
Energy to have a probability to create a strand break = 0
/clustering/algo/setEmax:
Energy to have a probability to create a strand break = 1
allow
* ClusterSBPoints: defines a cluster of strand break points
* CommandLineParser: defines a parser for command line control
as in other Geant4-DNA examples
* RunInitObserver: allows initializations at new run
(as in pdb4dna and microdosimetry)
* SBPoint: defines a class for point of energy deposition
----> Acknowledgments :
Ziad Francis for discussion about clustering algorithm.
...
SEE ALSO: README
