AIMS: This study retrospectively investigates the association between polygenic risk scores (PRS) derived from SNP clusters and glycaemic response to metformin in patients with newly diagnosed T2D. MATERIALS AND METHODS: Utilizing a dataset from the Taiwan Precision Medicine Initiative, we evaluated alterations in fasting glucose (FBG) and glycated haemoglobin (HbA1c) in individuals newly diagnosed with T2D who underwent metformin monotherapy for a duration of 6 months. Glycaemic responses between those in the bottom 20% of PRS (Q1) and the top 20% of PRS (Q5) for each of the SNP clusters and for the combination of two clusters were analysed. RESULTS: In responses to metformin monotherapy, significant differences of FBG levels were detected in Q1 as compared to Q5 in individuals of PRS derived from the cluster of beta-cell dysfunction with a positive association with proinsulin (Beta cell +PI) (p = 0.005) and the cluster of beta-cell dysfunction with a negative association with proinsulin (Beta cell -PI) (p = 0.003). Moreover, lower FBG levels on treatment were observed in those with both Q1 than those with both Q5 in the PRS derived from the two clusters of beta cell dysfunction (p = 0.002). Significantly reduced HbA1c values were documented in the Q1 in comparison to the Q5 within the cluster of Beta cell -PI (p = 0.002). CONCLUSION: These findings suggest that PRS derived from beta-cell dysfunction clusters may help predict glycaemic response to metformin and support the potential for genetically guided treatment in T2D.