Schematic overview of the mutation accumulation process and the mutation timing approach to separate dependent from independent events.

<p>(A) The occurrence of drivers is subject to hidden constraints, represented by a dependency structure (blue), whereas passengers are independent (orange). (B) Every cancer sample is an independent realization of the common underlying oncogenesis process. (C) Noisy cross-sectional mutation data from a set of tumor samples is the basis for discrimination of dependent from independent events. (D) Conditional mutation probabilities <i>P</i>_m(<i>k</i>) = Pr(mutation <i>m</i> has occurred given <i>k</i> or less mutations have occurred so far) of dependent mutations (blue) and (E) unconstrained mutations (orange) under the assumption of identical evolutionary rates have different characteristic shapes. (F) Schematic representation of a sigmoidal curve (black) used to approximate and the slope of this curve at the inflection point (red); this slope is used for ranking the genes or loci of interest.