Computational protocol: The Impact of 3′UTR Variants on Differential Expression of Candidate Cancer Susceptibility Genes

Similar protocols

Protocol publication

[…] 3′UTRs of the 39 genes mapping to Skts5 for which we did not have sequence data for all strains used in the linkage analyses were identified using the Ensembl database, builds 35–48. We designed PCR primers using Integrated DNA Technology's SciTools PrimerQuest web-based program [http://www.idtdna.com/Scitools/Applications/Primerquest; San Diego, CA]. PCR products were treated with Exo/SAP-IT (USB, Cleveland, OH) to remove single stranded DNA. Automated sequencing of PCR products was conducted on an ABI 3700 (Applied Biosystems/Life Technologies, Carlsbad, CA) by standard methods. Primers used for PCR were also used for the sequencing. Forward and reverse sequence reads were analyzed and compared using Lasergene/DNAstar 3.0 (DNASTAR, Madison, WI). The sequence traces were inspected visually whenever a nucleotide substitution was indicated. [...] 3′UTR polymorphisms that were observed only between NIH/Ola and SPRET/Outbred were evaluated to determine if they disrupted or introduced in silico predicted microRNA binding sites. Four programs were used: MiRanda (www.microRNA.org) , MicroInspector (http://bioinfo.uni-plovdiv.bg/microinspector/) , Patrocles finder (www.patrocles.org) and MicroSNiPer (http://cbdb.nimh.nih.gov/microsniper/getSeqByNM.php) . MiRanda allowed us to predict if our SNPs of interest disrupted miRNA binding sites in the mouse reference strain, which was highly similar to NIH/Ola 3′UTRs. The other three programs allow unique sequences to be analyzed for microRNA binding sites. For sites predicted by MicroInspector, we first considered those which had a predicted minimum free energy (MFE) of greater than −15 kcal/joule in one form and a MFE of less than −20 kcal/joule in the other form. When reevaluated, only those which were estimated to have a MFE of greater than −18 kcal/joule in one form and a MFE of −22 kcal/joule or less in the other form, as has been used previously for Mus musculus , were further tested experimentally. −22 kcal/J or less was considered strong binding and −18 kcal/J or higher was considered no or very weak binding. Patrocles and MicroSNiPer allow the comparison of two unique sequences for differences in predicted binding sites. MicroSNiPer requires SNPs to be entered into the prediction program, so this tool was unable to predict sites created or disrupted by insertions or deletions. Using MicroInspector, Patrocles, and MicroSNiPer, we picked candidate miRNAs that were predicted to bind to only the mouse strain (NIH/Ola or SPRET/Outbred), that demonstrated lower expression as measured by our 3′UTR luciferase assay expression and that contained a polymorphism in the miRNA binding site that fit with the mouse linkage results. [...] To refine our list of candidate microRNAs (miRNAs), we analyzed both the NIH/Ola and SPRET/Outbred forms of these binding sites in two minimum free energy (MFE) prediction tools, RNAhybrid (http://bibiserv.techfak.uni-bielefeld.de/rnahybrid/) , and RNAcofold (http://rna.tbi.univie.ac.at/cgi-bin/RNAcofold.cgi) . We further refined our list of candidate 3′UTRs to those whose SNPs were predicted to induce a 5 kcal/J or greater difference in miRNAbinding MFE between NIH/Ola and SPRET/Outbred by both prediction tools (). […]

Pipeline specifications

Software tools PrimerQuest, MiRanda, MicroInspector, MicroSNiPer
Databases microRNA.org Patrocles
Applications sRNA-seq analysis, qPCR
Organisms Mus musculus
Diseases Neoplasms, Skin Neoplasms