Introduction

Microarrays evolved from a highly specialized technique into a standard molecular biology method that is widely used for whole-genome gene expression profiling. However, many crucial steps are necessary to obtain meaningful results. Typically, multiple quality-control steps are employed to monitor sample and array quality throughout the entire array processing protocol and subsequent data analysis (1,2). One of the most important aspects of this quality control, however, is sample identity, that is, whether the expression profile recorded from an array actually derives from the indicated sample. There are several potential steps where a mix-up of samples may occur, for example column purification or array loading, where the entire reaction is transferred between different tubes. With the increasing size of microarray studies, it is important to ensure that each expression profile is assigned to the correct sample. Errors at this level almost certainly lead to erroneous results and can even cause a complete failure of the microarray study. In some cases, there might be some expression features that allow the confirmation of sample identity. Expression of sex-specific transcripts can be used, for example, to confirm the correct gender of the samples. In most cases, however, no such characteristic can be found. Very often this problem is addressed by the use of laboratory information management software (LIMS), which monitors each sample throughout the entire processing and keeps track of each step performed. Low- to medium- throughput laboratories, however, often have no access to LIMS resources. And even with LIMS support, there is always the factor of human error, which cannot be controlled completely. Thus, we developed a sample tracking system that utilizes probes already present on commercially available Affymetrix arrays to unambiguously correlate the recorded expression profile with the input sample RNA. A set of eight spike-in controls were generated which can be added to sample RNA in different combinations to generate an “on-chip identifier” which passes through the entire array processing protocol and results in a sample-specific hybridization pattern. This pattern can then be used to monitor whether each array was hybridized with the correct sample. The spike-in controls did not have any negative effect on RNA integrity or any detectable influence on the expression values of the remaining probes on the array and thus represent an inexpensive and easily adaptable system to guarantee high-quality microarray results.

Materials and methods

Construction of identifier constructs

Complementary oligonucleotides consisting of a tandem repeat of the sequence of the targeted probe fused to 20 adenosine residues were mixed at a final concentration of 10 µM and phosphorylated with polynucleotide kinase (Promega, Mannheim, Germany). The sequences of all oligonucleotides used are listed in Supplementary Table S1. After addition of NaCl to a final concentration of 60 mM, the oligonucleotides were heated to 96°C for 5 min and then cooled to 25°C with a ramp rate of 0.1°C/s in a thermal cycler to generate double-stranded fragments. These fragments also harbor 5′ overhanging ends compatible with BamHI and HindIII to allow insertion into pBlueskript SK II+ (Agilent, Waldbronn, Germany). Fragments for microarray hybridization were generated by PCR amplification of the insert region of the plasmids with 0.5 µM each standard M13 primer (F:GTAAACGACGGCCAGTG; R:GGAAACAGCTATG ACCATGA) with 2× Phusion HF PCR Master Mix (Finnzymes, Espoo, Finland) and 10 pg plasmid DNA as template. After 30 s of denaturation at 98°C, fragments were amplified with 35 cycles consisting of 10 s at 98°C, 30 s at 58°C, and 30 s at 72°C, and then underwent a final extension step at 72°C for 5 min. PCR products were purified using Qiagen PCR purification kit (Qiagen, Hilden, Germany) according to the manufacturer's protocol, and concentration was determined on a Nanodrop UV spectrophotometer (Thermo Scientific, Wilmington, DE, USA). Each fragment was incubated with 1 µg RNA from human embryonic kidney 293 (HEK) cells at a concentration of 1 pM for 4 h at 37°C. RNA was analyzed on an Agilent Bioanalyzer 2100 (Agilent, Waldbronn, Germany) to monitor effects of the fragments on RNA quality.

RNA extraction

RNA from HEK cells was extracted with Qiagen RNeasy Mini Kit, and rat tissue samples were homogenized in QIAzol reagent (Qiagen). After separation of organic and aqueous phases, the latter was loaded onto a Qiagen RNeasy Midi column to purify the RNA. RNA was quantitated with a Nanodrop UV spectrofluorometer and quality was monitored on an Agilent Bioanalyzer 2100.