Improvement in the Reproducibility and Accuracy of DNA Microarray Quantification by Optimizing Hybridization Conditions

Tao Han 1 2 3 Cathy D Melvin 1 2 3 Leming Shi 1 2 William S Branham 1 2 3 Carrie L Moland 1 2 3 P Scott Pine 0 1 Karol L Thompson 0 1 James C Fuscoe 1 2 3 0 Center for Drug Evaluation and Research, U.S. FDA , Silver Spring, MD 20993 , USA 1 from The Third Annual Conference of the MidSouth Computational Biology and Bioinformatics Society Baton Rouge , Louisiana. 2-4 March, 2006 2 Division of Systems Toxicology, National Center for Toxicological Research, U.S. FDA , Jefferson, AR 72079 , USA 3 Center for Functional Genomics, National Center for Toxicological Research, U.S. FDA , Jefferson, AR 72079 , USA Background: DNA microarrays, which have been increasingly used to monitor mRNA transcripts at a global level, can provide detailed insight into cellular processes involved in response to drugs and toxins. This is leading to new understandings of signaling networks that operate in the cell, and the molecular basis of diseases. Custom printed oligonucleotide arrays have proven to be an effective way to facilitate the applications of DNA microarray technology. A successful microarray experiment, however, involves many steps: well-designed oligonucleotide probes, printing, RNA extraction and labeling, hybridization, and imaging. Optimization is essential to generate reliable microarray data. Results: Hybridization and washing steps are crucial for a successful microarray experiment. By following the hybridization and washing conditions recommended by an oligonucleotide provider, it was found that the expression ratios were compressed greater than expected and data analysis revealed a high degree of non-specific binding. A series of experiments was conducted using rat mixed tissue RNA reference material (MTRRM) and other RNA samples to optimize the hybridization and washing conditions. The optimized hybridization and washing conditions greatly reduced the non-specific binding and improved the accuracy of spot intensity measurements. Conclusion: The results from the optimized hybridization and washing conditions greatly improved the reproducibility and accuracy of expression ratios. These experiments also suggested the importance of probe designs using better bioinformatics approaches and the need for common reference RNA samples for platform performance evaluation in order to fulfill the potential of DNA microarray technology. - Introduction DNA microarray has become the major tool to study global gene expression profiles in recent years [2,3]. Data from microarray experiments have been successfully used for establishing new pathways and identifying "signature" genes to differentiate cell types [4,5]. Because of the increased use of microarrays to analyze the gene transcriptional response, it is crucial to ensure the reproducibility, reliability, and accuracy of microarray data. DNA microarray is a very complex process involving many steps, such as probe design, array fabrication, RNA labeling, hybridization and washing, scanning, and data acquisition. Any missteps in the microarray process may lead to noise in the microarray experiment, which would adversely affect any conclusions drawn from the experiment. Various issues have been raised about the reliability and validity of microarray gene expression data [6-8]. For example, sub-optimally designed probes or incorrect probe annotations can lead to unreliable measurements in microarray experiments [9]. At a more fundamental level, a lack of consistency within and between different microarray platforms when the same RNA samples were tested has also been reported [6-8,10-12]. Such reports cast suspicion on microarray results and conclusions. Recent studies have shown, however, that carefully following established protocols, and using robust experimental designs and appropriate analytical methods can reduce the variability in microarray experiments and can result in much higher reproducibility and consistency [1316]. In addition, there are many technical issues that must be controlled in the fabrication and use of spotted microarrays that can have a dramatic impact on the quality of microarray data [17]. For example, intra-lab consistency can be improved by (1) the optimization of printing conditions such as relative humidity and buffer composition [18,19], (2) the optimization of purification procedures for RNA amplification and labeling [20,21], and (3) using consistent scanner power and voltage settings [22-24]. The fundamental basis of microarray technology is the specific binding (hybridization) of each probe to a labeled complementary target during the hybridization process [25]. The specificity of each oligonucleotide probe is associated with its melting temperature (Tm) and the salt concentration in the hybridization buffer. Welldesigned oligonucleotide sets should have a very narrow Tm range to ensure all the probes have very similar hybridization properties under the chosen hybridization condition. In this paper, we used tissue and mixed tissue RNA samples to assess the effect of hybridization and washing conditions on the microarray expression ratios. The reproducibility and accuracy (specificity) of microarray data were greatly improved with the optimized hybridization and washing conditions. These experiments also suggest that improvements in probe design will improve the reliability of microarray measurements and the ability to extract meaningful information from microarray data. Results Detection of non-specific binding under manufacturerrecommended hybridization condition To evaluate in-house printed Clontech 4 k rat oligonucleotide arrays, rat MTRRM (Mix1 and Mix2; see Materials and Methods) were labeled with cyanine dyes (Cy3 or Cy5) in a flip dye design and hybridized using the GlassHyb buffer at 50C for 1618 hours. All the slides were washed at room temperature with washing condition 1 (see Materials and Methods). Clontech probes were matched to tissue selective probes on Affymetrix RAE230A arrays based on Unigene ID [26]. Since the signal intensities of these tissue selective genes in the Mix1 and Mix2 samples fall into a wide range, they provide a simple and straightforward tool to use for platform evaluation. The results showed that the expression ratios of the tissueselective genes under the manufacturer-recommended conditions were greatly compressed compared to the theoretical input ratios (Figure 1; Table 1). It was difficult to distinguish any ratios different from 1 even though the input ratios were 4, 2, 1.5, and 1. The input ratios are easily observed with Affymetrix, Agilent, and Codelink microarray platforms [1]. Thus, this platform, under the Log2 Intensity of Mix1 fShFrcyiogabmturtierdMerizpTa1lotRitoRnoMfalMnodgix2w1saaisgnhndianlMginicxtoe2nn[ds1it]ieousnnsodefrtisnsoune-oseplteimctizvedgenes Scatterplot of log2 signal intensities of tissue selective genes from MTRRM Mix1 and Mix2 [1] under non-optimized hybridization and washing conditions. Parentheses indicate the total nu (...truncated)