Thorsten Will1, Volkhard Helms2. 1. Center for Bioinformatics and Graduate School of Computer Science, Saarland University, Saarbrücken, Germany. 2. Center for Bioinformatics and.
Abstract
UNLABELLED: Protein-protein interaction networks are an important component of modern systems biology. Yet, comparatively few efforts have been made to tailor their topology to the actual cellular condition being studied. Here, we present a network construction method that exploits expression data at the transcript-level and thus reveals alterations in protein connectivity not only caused by differential gene expression but also by alternative splicing. We achieved this by establishing a direct correspondence between individual protein interactions and underlying domain interactions in a complete but condition-unspecific protein interaction network. This knowledge was then used to infer the condition-specific presence of interactions from the dominant protein isoforms. When we compared contextualized interaction networks of matched normal and tumor samples in breast cancer, our transcript-based construction identified more significant alterations that affected proteins associated with cancerogenesis than a method that only uses gene expression data. The approach is provided as the user-friendly tool PPIXpress. AVAILABILITY AND IMPLEMENTATION: PPIXpress is available at https://sourceforge.net/projects/ppixpress/.
UNLABELLED: Protein-protein interaction networks are an important component of modern systems biology. Yet, comparatively few efforts have been made to tailor their topology to the actual cellular condition being studied. Here, we present a network construction method that exploits expression data at the transcript-level and thus reveals alterations in protein connectivity not only caused by differential gene expression but also by alternative splicing. We achieved this by establishing a direct correspondence between individual protein interactions and underlying domain interactions in a complete but condition-unspecific protein interaction network. This knowledge was then used to infer the condition-specific presence of interactions from the dominant protein isoforms. When we compared contextualized interaction networks of matched normal and tumor samples in breast cancer, our transcript-based construction identified more significant alterations that affected proteins associated with cancerogenesis than a method that only uses gene expression data. The approach is provided as the user-friendly tool PPIXpress. AVAILABILITY AND IMPLEMENTATION: PPIXpress is available at https://sourceforge.net/projects/ppixpress/.
Authors: Steven M Hill; Nicole K Nesser; Katie Johnson-Camacho; Mara Jeffress; Aimee Johnson; Chris Boniface; Simon E F Spencer; Yiling Lu; Laura M Heiser; Yancey Lawrence; Nupur T Pande; James E Korkola; Joe W Gray; Gordon B Mills; Sach Mukherjee; Paul T Spellman Journal: Cell Syst Date: 2016-12-22 Impact factor: 10.304
Authors: Khadija El Amrani; Gregorio Alanis-Lobato; Nancy Mah; Andreas Kurtz; Miguel A Andrade-Navarro Journal: PeerJ Date: 2019-05-27 Impact factor: 2.984
Authors: Xin Zhang; Christine S Gibhardt; Thorsten Will; Hedwig Stanisz; Christina Körbel; Miso Mitkovski; Ioana Stejerean; Sabrina Cappello; David Pacheu-Grau; Jan Dudek; Nasser Tahbaz; Lucas Mina; Thomas Simmen; Matthias W Laschke; Michael D Menger; Michael P Schön; Volkhard Helms; Barbara A Niemeyer; Peter Rehling; Adina Vultur; Ivan Bogeski Journal: EMBO J Date: 2019-07-15 Impact factor: 11.598