Identifying a minimal dose capable of eliciting a biological response is a fundamental issue in a number of scientific fields, including: drug development, signal transduction research, and environmental toxicology. Frequently, proliferation, viability, and other assays based on the cellular response to a treatment are used to assess the threshold dose for minimal activity. Here we propose a novel approach for identifying the effects of low dose treatments and pinpointing the threshold dose. Using microarrays, we examined the transcriptional response of a hormone responsive breast cancer cell line (MCF-7) stimulated with various concentrations of estrogen. Previous studies have focused on transcriptional responses to physiologically relevant concentrations of estrogen. However, relatively few studies have examined the transcriptional effects of concentrations below normal physiologic levels. These doses may not stimulate the expression of any genes or, alternatively, may regulate a different subset of genes that had not been previously characterized as estrogen responsive. We used gene expression profiling, coupled with a detailed analysis of replicates, to measure estrogen effects on many transcriptional targets and found that only physiologically relevant doses of estrogen (1 x 10(-10) M and higher) were capable of inducing a transcriptional response. This study demonstrates the utility of gene expression profiling as a means to identify concentrations that do not elicit a change in gene expression, or simply a No Observed Transcriptional Effect Level (NOTEL). The identification of a NOTEL for a given compound may be beneficial in several different scientific disciplines. For example, in the development of therapeutic drugs, a NOTEL could be used to identify doses of pharmaceutical compounds that are no longer effective at modulating the expression of biomarkers of efficacy.