Dividing somatic stem cells acquire DNA changes marking different clones. With time, clones can become large, either stochastically through neutral drift, or increased fitness and consequent selection. We present a simple, direct, and general approach that distinguishes between these two processes in normal somatic tissue in individuals. The method relies on single time point whole genome sequencing to study somatic mosaicism as tissues age. Using this method, we show that in human clonal hemopoiesis (CH), clones with CH driver mutations, that comprise a median of 24% of hematopoiesis originate decades before they are detected. They expand, through selection by a median of 26% per year. Overall, there is a 3-fold increased rate of stem cell division and an 8.6-fold increase in active long-term stem cells.One sentence summaryWe present an approach to distinguish genetic drift from selection, as subclones arise in somatic tissues, in individuals.