Introduction: The previous randomized phase 3 trial (SELECT BC) showed that S-1 as a first-line chemotherapy for metastatic breast cancer (MBC) is non-inferior to taxane with respect to overall survival. This study aimed to identify the usefulness of metabolism-related genes as predictive biomarkers for the response to S-1 compared with taxane using tumor tissue samples from the previous trial. Patients and Methods: In this SELECT BC-EURECA study, 147 patients with human epidermal growth factor 2 (HER2)-negative MBC who received either S-1 or taxane were evaluated. Formalin-fixed paraffin-embedded specimens were collected, and 14 genes involved in the pyrimidine metabolic pathway, estrogen receptor, progesterone receptor, HER2, Ki67, and beta-tubulin were measured using reverse transcription polymerase chain reaction in microdissected tumor specimens. The expression of each gene was categorized as low, intermediate, and high by tertile values. Results: Interaction tests to identify biomarkers for the response to S-1 compared with taxane, revealed the following as the top 3 biomarkers: RRM1 (P value = 0.24), GGH (P value = 0.25), and MTHFR (P value = 0.28). In the S-1 group, lower GGH and higher MTHFR expression were significantly correlated with better time to treatment failure. In the taxane group, there was no gene that was identified as a significant indicator of treatment failure. Conclusion: This biomarker analysis from SELECT BC did not identify any predictive biomarkers for the response to S-1 compared with taxane. Future studies with larger sample size and information on not only mRNA, but also protein and DNA for broad functional analyses are needed. Our previous randomized phase 3 trial showed that S-1 as a first-line chemotherapy for metastatic breast cancer is noninferior to taxane with respect to overall survival. In this study, we investigated the usefulness of metabolism-related genes as predictive biomarkers of treatment response to S-1 from the previous trial. We found no association between metabolic-related genes and treatment response to S-1.