Bauer, B & Gokhale, C. S.
Repeatability of evolution on epistatic landscapes
Scientific Reports, 5
“if we run the tape of life back from the start how likely is it that we will get the same outcome that we see around us today?” – Stephen Jay Gould
A nice overview by the first author Benedikt Bauer –
In many biological systems it is interesting to know how long it takes for a mutational process to happen. For example to be able to predict how long it takes for the bird flu to accumulate the necessary mutations to cross the interspecies barrier. Another example would be the time it takes for self cells to harvest the mutations necessary to turn into cancer cells. In our manuscript we present an algorithm on how to recursively compute the time distribution for processes in which the individuals proliferate independently of other individuals. By having the time distribution, it is possible to say up to which time the mutational process happens with a probability below a certain threshold. We go even step further in our manuscript and develop a procedure computing the time distribution for the single mutational pathways. This allows us to derive probabilities for the order in which the mutations are accumulated and ultimately this gives information about the subpopulations present in the system. In cancer, for example, subpopulations pose a high thread for a relapse after treatment: A subpopulation might not be targeted by normal chemotherapy and in the worst case only one additional mutations for cells of that subpopulation is needed to become cancerous. Therefore, by knowing which subpopulations are present it might be possible to develop therapies specific for those subpopulations. While there is no closed form solution for the time distributions, a direct computation of the mutational path probabilities would make it possible to interactively analyze the parameter space, even when not all required input, like mutation, birth, and death rates, are available, since long lasting simulations are not necessary anymore.