Figure 1

Individual-based model of clonal evolution with interactions. (A) Model architecture: division and death rates depend both on basal rates (b ₀ and d ₀) and on the effects of mutations on phenotype. Mutations affect clonal growth kinetics through a matrix of interactions, where b _ji and d _ji are the effects of clone j on clone i’s growth. (B) Evolution of the matrix of interactions: the matrix of interactions encompasses both cell-autonomous effects (b _ii and d _ii ) and non-cell-autonomous effects (b _∙,i and d _∙,i ). When new clones arise by mutation, the matrix is appended with new interaction coefficients r _∙,∙ which are drawn in normal distributions centered on interaction coefficients of the resident clone. (C) Probabilities of division and death: interactions coefficients are distinguished on whether they benefit or harm the recipient clone. They are then added to basal division and death rates to produce effective division and death rates. These effective rates are then used as parameters for exponential distribution in which division and death events are drawn. (D) Range of fitness effects: in a tight fitness landscape, the maxima of division and death rates are set to the range of mutation effects distribution, i.e., cell-autonomous and non-cell-autonomous effects similarly impact the division and death rates. In a wide fitness landscape, the maxima of division and death rates are higher than the maxima of mutation effects distribution, i.e., non-cell-autonomous effects can have distinct impacts on clonal growth rates from cell-autonomous effects’.