BACKGROUND: Pathogenic variants in susceptibility genes lead to increased breast cancer risk. METHODS: To identify coding variants associated with breast cancer risk, we conducted whole exome sequencing in genomic DNA samples from 831 breast cancer cases and 839 controls of Chinese women. We also genotyped samples including 4,580 breast cancer cases and 6,695 controls using whole exome-chip arrays. We further performed a replication study using a Multi-Ethnic Global Array (MEGA) in samples from 1,793 breast cases and 2,059 controls. A single marker analysis was performed using the Fisher Exact Test. RESULTS: We identified a missense variant (rs139379666, P2974L; AF = 0.09% for breast cancer cases, but none for controls) in the ATM gene for breast cancer risk using combing data from 7,204 breast cancer cases and 9,593 controls (P = 1.7 × 10-5). To investigate the functionality of the variant, we first silenced ATM and then transfected the overexpression vectors of ATM containing the risk alleles (TT) or reference alleles (CC) of the variant in U2OS and breast cancer SK-BR3 cells, respectively. Our results showed that compared to the reference allele, the risk allele significantly disrupts the activity of homologous recombination (HR)-mediated double-strand breaks (DSB) repair efficiency. Our results further showed that the risk allele may play a defected regulation role in the activity of the ATM structure. CONCLUSIONS: Our findings identified a novel mutation that disrupts ATM function, conferring to breast cancer risk. IMPACT: Functional investigation of genetic association findings to discover a pathogenic variant for breast cancer risk.