Page History

...

The Flux Simulator uses a 3-step algorithm to tokenize a molecule; first, geometry  and the number  of fragments that are obtained from the molecule are determined. We found empirically that parameter d depends logarithmically on , the length of the molecule that is fragmented . The number of fragments produced from a specific RNA molecule is determined by , where  is the expectancy of the most abundant fragment size, computed from h and the gamma-function  of :

Second,  breakpoints are sampled uniformly from the interval [0;1[, resulting in relative length fractions  for all all  fragments. Third, relative fragment sizes  are transformed from unit space to sizes  that follow a Weibull distribution of shape d shape  by:

where  is a constant of the transformation to ensure that the sizes of the  fragments sum up exactly to the given molecule length. Latter transformation produces a slightly distorted Weibull distribution for the sizes , however the deviation is sufficiently small to be neglected in our applications.