The story of how, within just a few weeks, a scientist can uncover plausible biomarkers for a complex disease using IPA® for in silico research. We invite you to register for a webinar to hear Dr. Billaud present this in silico study live on September 21 or to listen to a pre-recorded video recording of his presentation.
By Jean-Noel Billaud, Associate Staff Scientist, Ingenuity. September 20, 2011
(View Part I, Part II, or Part III)
Part IV: Combining a microRNA dataset with mRNA expression data using the IPA microRNA Target Filter
The advantage of analyzing different aspects of the biology in IPA is exemplified by the new microRNA Target Filter. This new feature was released early in 2011 as part of IPA 9.0. The microRNA Target Filter associates microRNAs from a dataset with experimentally observed and predicted mRNA targets and allows researchers to prioritize targets using their experimental results coupled with rich information that IPA exposes from the Ingenuity® Knowledge Base throughout the biological analysis in IPA.
I took advantage of this capability by uploading a recent microRNA microarray analysis published by Jung et al. in the International Journal of Cancer. In this study (which I’ll refer to as Study 2), the authors screened the expression of 470 miRNAs by microarray analysis in 24 matched pairs of histologically confirmed tumor tissue and normal adjacent tissue. They successfully validated their results using RT-qPCR analyses of 76 of these tissue pairs. They then analyzed the associations of miRNA expression with clinico-pathological data and evaluated the potential of miRNAs as diagnostic and/or prognostic markers for PCa. They “identified 15 differentially expressed miRNAs (10 downregulated and 5 upregulated), and expression 5 of these miRNAs correlated with Gleason score, a grading score based on microscopic appearance of the tumor, or pathological tumor stage.” IPA successfully mapped 14 of the 15 miRNAs into IPA. The microRNA Target Filter in IPA indicated that 13 out of the 14 miRNAs are associated with 8009 mRNAs either experimentally determined (TarBase) or predicted (Target Scan). The microRNA Target Filter enabled filtering on both the source and confidence of the microRNA-mRNA target relationships. I selected two levels of stringency of relationships between the miRNA and mRNA (to limit targets to those that were experimentally observed or predicted with high confidence. Then I used the previously defined list of SRF and androgens-dependent genes (158 genes) to prioritize the targets having relationships with the 13 miRNAs from the miRNA Study 2. I also applied the filtered to select inversely expressed target pairings to prioritize miRNA-mRNA target relationship based on expression changes. This filtering resulted in narrowing down to 12 miRNAs that could be targeting 35 mRNAs of the 158 possible from Study 1 (See Figure 6).
In summary, the microRNA Target Filter allowed me to successfully integrate two independent studies in IPA and identify convergent and relevant key molecular aspects that may be useful for identifying potential biomarker for PCa.
Next week, I will show how this in silico research will culminate in identifying new hypothetical biomarkers, backed by several different lines of published evidence, that warrant further study as prostate cancer biomarkers.
Click here to go to the Ingenuity.com home page.