Adaptor-tagged competitive polymerase chain reaction: amplification bias and quantified gene expression levels Hiroko Kita-Matsuo a,1 , Naoto Yukinawa b , Ryo Matoba a , Sakae Saito a , Shigeyuki Oba b , Shin Ishii b , Kikuya Kato a, * a Taisho Laboratory of Functional Genomics, Nara Institute of Science Technology, 8916-5 Takayama, Ikoma, Nara 630-0101, Japan b Laboratory of Theoretical Life Science, Nara Institute of Science and Technology, 8916-5 Takayama, Ikoma, Nara 630-0101, Japan Received 10 August 2004 Available online 21 January 2005 Abstract Adaptor-tagged competitive polymerase chain reaction (ATAC-PCR) is an advanced version of quantitative competitive PCR characterized by the addition of unique adaptors to different cDNA samples. It is currently the only quantitative PCR technique that enables large-scale gene expression analysis. Multiplex application of ATAC-PCR employs seven adaptors, two or three of which are used as controls to generate a calibration curve. The characteristics of the ATAC-PCR method for large-scale data pro- duction, including any adaptor- and gene-dependent amplification biases, were evaluated by using this method to analyze the expres- sion of 384 mouse brain genes. Short adaptors tended to amplify at higher efficiency than did long adaptors. The population of genes with a high amplification bias increased with the use of short adaptors. Subtracting the median value of all adaptor-dependent biases could reduce this bias; the majority of genes displayed a small gene-dependent bias, which facilitated reliable quantification. We modified ATAC-PCR to estimate molecular numbers of transcripts by introducing synthetic standards. This modification demon- strated that gene expression levels in mammalian cells are varied over seven orders of magnitude. Ó 2004 Elsevier Inc. All rights reserved. Keywords: ATAC-PCR; Gene expression profiling; RT-PCR; PC12 There are two types of technical approaches to gene expression profiling. One involves random sample sequencing of a pool of cloned genes to count the num- ber of RNA molecules present. Novel techniques using this approach include serial analysis of gene expression (SAGE) 2 [1] and massively parallel signature sequencing (MPSS) [2]. These techniques are appropriate for many purposes, including characterization of tissue specificity. However, due to the presence of experimental artifacts and low throughput, they are not useful for analysis of a large number of samples such as the analysis of human cancer tissues to identify diagnostic genes. In such cases, the second approach—the large-scale appli- cation of conventional quantitation methods—would be applicable. One technique based on this approach, the use of DNA microarrays [3], is currently the most 0003-2697/$ - see front matter Ó 2004 Elsevier Inc. All rights reserved. doi:10.1016/j.ab.2004.11.014 * Corresponding author. Present address: Osaka Medical Center for Cancer and Cardiovascular Diseases, 1-3-2 Nakamichi, Higashinari-ku, Osaka 537-8511, Japan. Fax: +81 66973 1209. E-mail address: katou-ki@mc.pref.osaka.jp (K. Kato). 1 Present address: The Burnham Institute, 10901 North Torrey Pines Road, La Jolla, CA 92037, USA. 2 Abbreviations used: SAGE, serial analysis of gene expression; MPSS, massively parallel signature sequencing; PCR, polymerase chain reaction; RT, reverse transcription; ATAC-PCR, adaptor-tagged competitive PCR; FAM, 6-carboxyfluorescein; HEX, hexachloro-6-carboxyfluorescein; VIC, Applied Biosystems proprietary dye; EDTA, ethylenediamine tetraacetic acid; MB-x, MB adaptor; DTT, dithiothreitol; ATP, adenosine triphosphate; DW, distilled water; BSA, bovine serum albumin; RMSE, root mean square error. ANALYTICAL BIOCHEMISTRY Analytical Biochemistry 339 (2005) 15–28 www.elsevier.com/locate/yabio