| Summary: | The mechanism of nonsense-mediated decay (NMD) selectively degrades mRNAs carrying a premature translation-termination codon and regulates the abundance of a large number of physiological mRNAs that encode full-length proteins. Although this complex process has been extensively studied along the years, the interactions and connectivity among NMD players is not completely understood. Additionally, some NMD mechanistical aspects suggest missing roles that can be played by proteins still not reported as involved in this pathway. To tackle this hypothesis, we developed a bioinformatic network-based approach to predict new proteins involved in NMD. Our approach consists in performing several queries to different types of publicly available data, in order to explore the ability of proteins to bridge related processes, while integrating data regarding protein-protein interactions, co-expression and co-regulation. We found that known NMD-factors have physical, regulatory and co-expression interaction signatures with related processes (mRNA translation, mRNA splicing, mRNA degradation and mRNA transport), which can be used to distinguish them from other proteins. We computed a scoring algorithm to rank NMD-neighbors according to the similarity to these signatures, generating a list of NMD candidates, that we aim to validate experimentally. Interestingly, some candidates were recently studied in NMD context and showed promising results. Furthermore, a cross-validation analysis indicated the robustness of the predictions provided by our method. On the road to developing a tool to apply this approach to other biological processes, we observed good cross-validations results for other RNA-related processes, suggesting this method’s usefulness in the RNA research area.
|
|---|