Mathematics Colloquia and Seminars
Return to Colloquia & Seminar listing
Predicting cancer risk from earth and space radiation using DNA damage clustering phenotypeMathematical Biology
A.C. Heuskin, A. Osseiran, J. Tang and S.V. Costes*
The modeling component of our NASA Specialized Center of Research (NSCOR) has developed several approaches to simulate the response of complex tissues to ionizing radiation over the years. Our work has benefited from the usage of agent-based models, a stochastic approach simulating life and emerging properties of complex interacting entities. This is done by imposing behavioral, signaling and physical rules on individual cells and their microenvionment. We introduce in this work a simpler approach using automata instead of agent-based model. Automata simulations are faster to compute, allowing the generation of larger set of predictions. On the other hand, automata are also more limited in the type of complexity they can model. Automata are used here to predict tumor incidence of human cohorts exposed to various radiation types.
We first implemented the multi-stage clonal expansion model to simulate tumor incidence arising spontaneously in human population due to random mutations. The simplest model we tested so far assumes that tumor arises via only two consecutive events: initiation and conversion. Each simulated person is a monolayer of 10,000 cells (referred as “tissue”) tracked from the age of 20 to 90 years old and where cells divide only when neighboring cells have died. At each division, mutation is possible with a set probability value. As a cell is more mutated, it becomes more unstable (mutation and death rates both increase). Note that the model has been set to allow more stages before tumor arises if necessary (i.e. multi-stage). Once a cell has become fully transformed (reached a set number of mutations), it can invade neighboring cells with an adjustable parameter for invasion efficiency. Parameter sweep was conducted to identify the rates for mutation and death frequency that lead to the closest predictions for breast cancer incidence in non-exposed normal human populations.
Next, radiation effects are simulated by generating a pulse of mutation and death whose dose and LET dependence are derived from our recently published DNA damage clustering phenotype. In addition, there are also non-targeted effects (NTE) of radiation, characterized by non-linear perturbation of the full organism. NTE is model by a switch-like function with a dose threshold in the cGy range, no LET dependence and adjustable duration. Briefly, NTE are being simulated by increasing transiently the transformation rate, the cell death rate, and the mutation rate at different dose threshold. The duration of these transient states and their dose thresholds are currently being investigated in silico so that simulations match incidence and age dependence observed from breast epidemiological data following X-ray exposure. Cosmic radiation effects remain to be simulated and will reflect the impact of the dose heterogeneity at the sub-micron level using a simple core/penumbra model. This work will lead to a set of theoretical relative biological effectiveness (RBE) for tumor incidence in in silico populations and can be used by NASA in their risk model for missions in space.
Please let Mariel Vazquez (email@example.com) know if you would like to meet with the speaker, or join him for dinner.