Intense genome sequencing of Pseudomonas aeruginosa isolates from cystic fibrosis (CF) airways has shown inefficient eradication of the infecting bacteria, as well as previously undocumented patient-to-patient transmission of adapted clones. However, genome sequencing has limited potential as a predictor of chronic infection and of the adaptive state during infection, and thus there is increasing interest in linking phenotypic traits to the genome sequences. Phenotypic information ranges from genome-wide transcriptomic analysis of patient samples to determination of more specific traits associated with metabolic changes, stress responses, antibiotic resistance and tolerance, biofilm formation and slow growth. Environmental conditions in the CF lung shape both genetic and phenotypic changes of P. aeruginosa during infection. In this Review, we discuss the adaptive and evolutionary trajectories that lead to early diversification and late convergence, which enable P. aeruginosa to succeed in this niche, and we point out how knowledge of these biological features may be used to guide diagnosis and therapy.