Background: Due to relatively high stability in serum/plasma, urine and other biofluids, miRNA profiling is currently being established as an exploratory tool to identify noninvasive biomarkers for human disease. Because of the small size of mature miRNAs, the high degree of homology between miRNA family members, and the low abundance of miRNAs in biofluids, miRNA expression profiling is technically challenging. Several Seq-based platforms have emerged and are attractive because of high sensitivity and whole miRNome survey but a rigorous platform evaluation is necessary.
Methods: We expand the miRQC study of Mestdagh et al. to evaluate newly advanced Seq-based platforms QIAGEN QIAseq, HTG EdgeSeq, Exiqon mirSeq with Taqman based Exiqon array and focus on more challenging biofluid miRNA profiling for objectively assess the performance of titration response, reproducibility, specificity, sensitivity, differential expression, and accuracy. We prepared identical 20 samples at Phase I and additional 9 samples at phase II for accuracy for QIAseq only.
Results: For specificity assessment, we spiked in 4 different miRNAs in liver and phage samples with 1 or 2 nucleotide differences. QIAseq ranked the highest for specific by cross reactivity calculation. The detection rate in serum RNA were much more variable among platforms with up to 6 fold differences. QIAseq shows the highest, while Exiqon mirSeq is at the low end for detection rate. We calculated raw UMI from phage spike in samples, the difference between the expected versus observed FC are range from 0.98 to 1.14 and demonstrated high accuracy for QIAseq platform. The reproducibility appears to be comparable with >80% corrections among serum samples. We applied the product rank sum method and identified 443 out of 2524 miRNAs as differentially expressed for QIAseq, 345 out of 2096 for EdgeSeq and 175 out of 794 for mirSeq at 0.05 cut-off p-value. For titration response, we computed the % of correctly titrating miRNAs as the function of Fold-change (FC) from Universal Human miRNA Reference RNA and Human Brain RNA. QIAseq is close to Exiqon Taman platform on the top performance with > 90% of the top half FC-ranked miRNAs titrated correctly.
Conclusion: QIAGEN QIAseq platform performed better in specificity, serum miRNAs detection rate, titration response among the three Seq-based platforms and outperforms the bench mark Exiqon Taqman, while QIAGEN QIAseq platform performed comparably in reproducibility, differential analysis among the four platforms. QIAseq also demonstrates high accuracy at phase II study and has been chosen the platform for biofluid miRNA clinical studies.
References: Ø P. Mestdagh et al “Evaluation of quantitative miRNA expression platforms in the microRNRNA quality control (miRQC) study” Nature Methods doi:10.1038/nmeth.3014
Citation Format: Beihong Hu, Desai H. Keyur, Zhenhao QI. Evaluation of seq-based biofluid miRNA profiling platforms [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2018; 2018 Apr 14-18; Chicago, IL. Philadelphia (PA): AACR; Cancer Res 2018;78(13 Suppl):Abstract nr 5391.