STUDY QUESTION: Do circadian genes exhibit an altered profile in peripheral blood mononuclear cells (PBMCs) of polycystic ovary syndrome (PCOS) patients and do they have a potential role in androgen excess? SUMMARY ANSWER: Our findings revealed that an impaired circadian clock could hamper the regulation of peripheral steroid metabolism in PCOS women. WHAT IS KNOWN ALREADY: PCOS patients exhibit features of metabolic syndrome. Circadian rhythm disruption is involved in the development of metabolic diseases and subfertility. An association between shift work and the incidence of PCOS in females was recently reported. STUDY DESIGN, SIZE, DURATION: This is a retrospective case-referent study in which peripheral blood samples were obtained from 101 control and 101 PCOS subjects. PCOS diagnoses were based on Rotterdam Consensus criteria. PARTICIPANTS/MATERIALS, SETTING, METHODS: This study comprised 101 women with PCOS and 101 control volunteers, as well as Swiss albino mice treated with dehydroepiandrosterone (DHEA) to induce PCOS development. Gene expression analyses of circadian and steroidogenesis genes in human PBMC and mice ovaries and blood were executed by quantitative real-time PCR. MAIN RESULTS AND THE ROLE OF CHANCE: We observed aberrant expression of peripheral circadian clock genes in PCOS, with a significant reduction in the core clock genes, circadian locomotor output cycles kaput (CLOCK) (P ≤ 0.00001), brain and muscle ARNT-like 1 (BMAL1) (P ≤ 0.00001) and NPAS2 (P ≤ 0.001), and upregulation of their negative feedback loop genes, CRY1 (P ≤ 0.00003), CRY2 (P ≤ 0.00006), PER1 (P ≤ 0.003), PER2 (P ≤ 0.002), DEC1 (P ≤ 0.0001) and DEC2 (P ≤ 0.00005). Transcript levels of an additional feedback loop regulating BMAL1 showed varied expression, with reduced RORA (P ≤ 0.008) and increased NR1D1 (P ≤ 0.02) in PCOS patients in comparison with the control group. We also demonstrated the expression pattern of clock genes in PBMCs of PCOS women at three different time points. PCOS patients also exhibited increased mRNA levels of steroidogenic enzymes like StAR (P ≤ 0.0005), CYP17A1 (P ≤ 0.005), SRD5A1 (P ≤ 0.00006) and SRD5A2 (P ≤ 0.009). Knockdown of CLOCK/BMAL1 in PBMCs resulted in a significant reduction in estradiol production, by reducing CYP19A1 and a significant increase in dihydrotestosterone production, by upregulating SRD5A1 and SRD5A2 in PBMCs. Our data also showed that CYP17A1 as a direct CLOCK-BMAL1 target in PBMCs. Phenotypic classification of PCOS subgroups showed a higher variation in expression of clock genes and steroidogenesis genes with phenotype A of PCOS. In alignment with the above results, altered expression of ovarian core clock genes (Clock, Bmal1 and Per2) was found in DHEA-treated PCOS mice. The expression of peripheral blood core clock genes in DHEA-induced PCOS mice was less robust and showed a loss of periodicity in comparison with that of control mice. LARGE SCALE DATA: N/A. LIMITATIONS, REASONS FOR CAUTION: We could not evaluate the circadian oscillation of clock genes and clock-controlled genes over a 24-h period in the peripheral blood of control versus PCOS subjects. Additionally, circadian genes in the ovaries of PCOS women could not be evaluated due to limitations in sample availability, hence we employed the androgen excess mouse model of PCOS for ovarian circadian assessment. Clock genes were assessed in the whole ovary of the androgen excess mouse model of PCOS rather than in granulosa cells, which is another limitation of the present work. WIDER IMPLICATIONS OF THE FINDINGS: Our observations suggest that the biological clock is one of the contributing factors in androgen excess in PCOS, owing to its potential role in modulating peripheral androgen metabolism. Considering the increasing prevalence of PCOS and the rising frequency of delayed circadian rhythms and insufficient sleep among women, our study emphasizes the potential in modulating circadian rhythm as an important strategy in PCOS management, and further research on this aspect is highly warranted. STUDY FUNDING/COMPETING INTEREST(S): This work was supported by the RGCB-DBT Core Funds and a grant (#BT/PR29996/MED/97/472/2020) from the Department of Biotechnology (DBT), India, to M.L. B.S.J. was supported by a DST/INSPIRE Fellowship/2015/IF150361 and M.B.K. was supported by the Research Fellowship from Council of Scientific \& Industrial Research (CSIR) (10.2(5)/2007(ii).E.U.II). The authors declare no competing interests. TRIAL REGISTRATION NUMBER: N/A.