Peripheral neuropathy is an adverse event of taxane treatment that is related both to the patient's cumulative drug exposure and their inherent sensitivity to neurotoxicity. Discovery and validation of genetic loci that determine neuropathy risk is an important first step towards individualization of taxane treatment with the ultimate goal of maximizing treatment efficacy and minimizing the risk of severe adverse events. Paclitaxel exposure is regulated by enzymes and transporters that have common variants known to influence protein expression or activity. Paclitaxel is primarily metabolized by the CYP2C8 enzyme, and prior research from our group and others suggests that patients who carry a common low-activity variant, CYP2C8*3, may be at increased risk of neuropathy. Using a cohort of paclitaxel-treated breast cancer patients, I was able to confirm the association between CYP2C8*3 and increased risk of paclitaxel-induced peripheral neuropathy. I then attempted to use a genotyping platform that interrogates thousands of variants in hundreds of genes relevant to drug metabolism, elimination, and transport to identify polymorphisms that influence risk of neurotoxicity after accounting for the CYP2C8*3 variant. Surprisingly, I discovered a polymorphism in a gene not thought to be relevant to paclitaxel pharmacokinetics, ABCG1, which was associated with neuropathy risk. Less is known about the clinical or genetic factors that modulate docetaxel-induced neuropathy risk. I performed genome-wide association in a large cohort of docetaxel-treated patients to discover genetic loci that modulate risk of neuropathy. I discovered several candidates, one of which was an intergenic polymorphism that surpassed genome-wide significance after adjustment for relevant clinical covariates. I then attempted, unsuccessfully, to replicate these discoveries in independent cohorts of taxane-treated patients. This inability to replicate indicates that either the associations of these variants are limited to the cohort in which they were discovered or that they were merely spurious discoveries. Replication should be attempted in independent patient cohorts that are more similar to those in which these discoveries were made to validate the influence of these variants on neuropathy risk, enabling translation into routine clinical practice.