MEMPHIS, Tenn. – Several research groups under the auspices of the Association of Biomolecular Resource Facilities are comparing how different microarray platforms stack up when it comes to microRNA analysis.
According to preliminary results of the ongoing study that were presented at ABRF's annual meeting here today, miRNA arrays from Affymetrix, Illumina, Agilent, and Exiqon all show different strengths and shortcomings regarding sensitivity and reproducibility.
The study, conducted jointly by ABRF's Microarrays Research Group and DNA Sequencing Research Group, compared the four array platforms to Applied Biosystems' TaqMan miRNA assay. The participants are also comparing the results to miRNA data from the Illumina Genome Analyzer, but that part of the project has not yet been completed.
In a talk outlining the preliminary results, Susan Hester of the US Environmental Protection Agency noted that the research groups aimed to assess the effectiveness of second-generation sequencing and microarrays for miRNA expression profiling.
Participating sites analyzed brain and liver RNA and ran three replicates on each array platform. The study used Applied Biosystems' RT-PCR platform as a baseline for comparing the accuracy of the different array technologies.
The study evaluated a number of parameters, including the number of miRNAs reproducibly detected with each platform, the number of miRNAs that were differentially expressed in brain and liver tissue by an arbitrary cutoff of more than two-fold, and the reproducibility within and across platforms.
The Illumina platform was the most sensitive, detecting 1,146 miRNAs, while the TaqMan assay detected 673. The Affy, Agilent, and Exiqon platforms detected 847, 723, and 732 miRNAs, respectively.
The Illumina platform was also the most sensitive when it came to differential expression, picking up 394 miRNAs with a fold change of more than two. The lowest in that category was the Affy platform, which measured 292 differentially expressed miRNAs with the cutoff fold change of two.
The Agilent platform had the best correlation with the TaqMan platform, however, and also demonstrated the highest sensitivity for miRNAs that were measured on all platforms, as well as the highest reproducibility of results generated by other platforms.
In terms of overall technical reproducibility, Affy came out ahead, according to the preliminary results, while the Exiqon and Agilent platforms showed the best cross-platform correlation of results.
While the analysis of the data is still ongoing, Hester said that the early results indicate that all platforms show both "strengths and underperformance on some sensitivity parameters."
One possibility for the variation in results, she said, could be differences in miRNA probes across the platforms, though she stressed that is just conjecture at this point.
Peter Schweitzer of Cornell University discussed his group's work to generate miRNA data from the Illumina Genome Analyzer for the study, but noted that so far the group has only sequenced the brain sample.
Schweizer said that the sequencer detected 359 miRNAs in that sample, and that an initial comparison to the Agilent array results indicates "good correlation," but stressed that the results are preliminary. He did not provide a timeline for the expected completion date of the sequencing component of the study.