Background— Few rare variants in atrial fibrillation (AF)–associated genes have been functionally characterized to identify a causal relationship between these variants and development of AF. We here sought to determine the clinical effect of rare variants in AF-associated genes in patients with lone AF and characterized these variants electrophysiologically and bioinformatically. Methods and Results— We screened all coding regions in 12 AF-associated genes in 90 patients with lone AF, with an onset of 47±11 years (66 men; mean age, 56±13 years) by high-resolution melting curve analysis and DNA sequencing. The potassium and sodium currents were analyzed using whole-cell patch clamping. In addition to using 4 individual in silico prediction tools, we extended those predictions to an integrated tool (Combined Annotation Dependent Depletion). We identified 7 rare variants in KCNA5 , KCNQ1 , KCNH2 , SCN5A , and SCN1B genes in 8 patients: 2 of 8 probands had a family history of AF. Electrophysiological studies revealed that 2 variants showed a loss-of-function, and 4 variants showed a gain-of-function. Five of 6 variants with electrophysiological abnormalities were predicted as pathogenic by Combined Annotation Dependent Depletion scores. Conclusions— In our cohort of patients with lone AF, 7 rare variants in cardiac ion channels were identified in 8 probands. A combination of electrophysiological studies and in silico predictions showed that these variants could contribute to the development of lone AF, although further in vivo study is necessary to confirm these results. More than half of AF-associated rare variants showed gain-of-function behavior, which may be targeted using genotype-specific pharmacological therapy.