Abstract A housekeeping gene is described as a gene that is stably expressed across tissues under normal conditions. These genes are extensively used in gene expression studies, particularly those involving quantitative real-time PCR, as a means to control for overall RNA abundance. Previous studies have reported extensive circadian variation at the transcriptional level, with as many as 10% of all expressed transcripts cycling over the course of a day. To determine the influence of circadian fluctuation on housekeeping gene stability, and the corresponding implications for estimates of gene expression, we evaluated time-course RNA-seq data from primary human mammary epithelial cells. These data demonstrate substantial time-dependent variability in the expression of several commonly used housekeeping genes. Of the 12 housekeeping genes evaluated, 8 had greater than 15% coefficient of variation (CV) with b-actin and GAPDH having CVs of 33.9% and 36.9%, respectively. In order to assess the impact of housekeeping gene variability on expression estimates of target genes which may use housekeeping transcript levels for the purposes of normalization, we stratified the 12 housekeeping genes into three categories: high variability, moderate variability, and low variability, and analyzed the effect of time-dependent fluctuations in several clinically-relevant targets, including ERBB2 and the genes used in a breast cancer panel to predict likelihood of recurrence. ERBB2, a proto-oncogene, is over-expressed in 15-30% of breast tumors and is associated with increased disease recurrence and a poor prognosis. Evaluating ERBB2 expression in our RNA-Seq data, which is normalized internally using both transcript length and total reads, we find that expression of ERBB2 is relatively stable over time (average FPKM = 23.28, SD= 2.91). However, when expression is normalized to genes in the high variability housekeeping set, expression fluctuates by as much as 50% above or below the mean. In contrast, normalization to genes in the low variability set provides an accurate depiction of expression across all time points. We also evaluated all genes in a commonly used clinical assay for predicting recurrence in newly diagnosed, early stage breast cancer patients. We find that a majority of the 16 genes included in the assay vary considerably over time when normalized to the 5 housekeeping genes identified by the assay developer, which could have important implications for result interpretation. Our RNA-Seq time-course data suggests that there is circadian variability in housekeeping gene expression and that it is important to consider circadian fluctuations during the experimental design phase in order to identify appropriate genes for RNA content normalization. Citation Format: Tracy L. Peters, Elizabeth J. Ferree, Yaou Sheng, Aaron E. Hoffman. Circadian fluctuations in “housekeeping” gene expression measured by RNA-seq. [abstract]. In: Proceedings of the 105th Annual Meeting of the American Association for Cancer Research; 2014 Apr 5-9; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2014;74(19 Suppl):Abstract nr 3403. doi:10.1158/1538-7445.AM2014-3403