KEPAMOD · Knowledge exchange in processing and analysis of multi-omic data

Knowledge exchange in processing and analysis of multi-omic data

The primary aim of KEPAMOD was to exchange knowledge in the processing and analysis of multi-omic data across research institutes via a host of staff exchanges for both experienced and young researchers. The project was split into three work packages. The first work package, between China (PICB) and Denmark (AU) had three main objectives. 1. PICB gain understanding for the proteomics techniques & methodologies. 2. AU gain understanding for the functional genomic techniques methodologies 3. Selection of compatible multi-omic datasets for future research. In order to achieve Objective 1, an experienced researcher from PICB spent 1 month in Denmark, gaining an understanding for proteomics by the characterization of AQP2 using proteomic techniques. To give some detail, the AQP2 protein is mainly known as a water channel, expressed in kidney collecting ducts, which contributes critically to water homeostasis in mammals. Furthermore AQP2 is not only a water channel but also an integrin-binding membrane protein that promotes renal epithelial cell migration and morphogenesis. During their exchange, an epitelial cell line model expressing AQP2 was used in order to obtain high amounts of the protein. Then AQP2 was isolated, purified and prepared for Mass Spectrometry (MS) analysis by the ER from PICB. Already processed data was then used to learn about the theory of MS and this knowledge transferred back to PICB. For the return exchange, a young researcher who was undertaking a PhD investigating the importance of the hormone, aldosterone, which is highly important for correct regulation of the blood pressure, travelled to China to learn how to analyse RNA sequencing data, which produces a massive quantity of complex data. At PICB, the young researcher gained skills within bioinformatics and programming, which are difficult to obtain without supervision from skilled researchers. Following, he used this knowledge to analyse his own data. The data analysis required super computers, which are found at PICB, and the analysis was not possible to conduct at his home university, Aarhus University. Finally, the PhD student has through the exchange program experienced being a part of an international strong research team, important in becoming an experienced, successful researcher. Work package 2 had 3 objectives: 1. EU scientists to become familiar with current multi-omic data integration techniques 2. Develop a model that will minimise the error from different omic data 3. Apply the integration technique to current datasets for testing The objectives of this work package were met and the results of the package have been published in Bioinformatics (Dent, Yang and Nardini, 2013 and Dent, Devescovi, Li, Di Lena, Lu, Liu, & Nardini, 2015), the leading journal in its field. The work package involved three months of exchange between both China and UK. Through a series of meetings, scientists in the EU and China determined that one of the major issues with current multi-omic data integration techniques is that they rely on multiple manual steps to transform data from their static representation to a format that can be used in simulation models. As part of WP2, we therefore developed a freely available application that allows for the transfer of multi-omic data from commonly used network representation software, such as CellDesigner and Cytoscape into BioLayout Express, software developed for the dynamic simulation of molecular interaction networks. The introduction of an automated step immediately reduces error. The application was tested on a large network that have been manually transferred and analysed in by the two groups in collaboration (Dent and Nardini, 2013), as well as on simple networks so that the error rates could be analysed. By testing the software on simple networks, it is easier to identify errors and to minimize them. On tests over larger networks, we found that there is a threshold number simulation runs, after which accuracy levels do not improve, the results of which are published in (Dent, Devescovi, Li, Di Lena, Lu, Liu, & Nardini, 2015). The work from work package two was also presented by the Researcher from the UK, in China and in addition to planned dissemination of results, has also been presented at a workshop in Germany by the Chinese scientist. Work package three had 2 main objectives: 1. PICB to complete series of seminars on translational research, at CNR. 2. Extension of a molecular interaction network for RA, using integrated multi-omic data For the first objective, an experienced researcher from China gave seminars to frame the area of research in which PICB aimed to apply the methods developed during the exchange, focusing on a) Complex networks with applications to rheumatoid arthritis and , b) Mechanotransduction - explaining how maps and simulations allow us to investigate the effects of therapeutic manipulations in inflammatory disease models. As planned the seminars related to the issues and challenges encountered in multiomic integration, and results obtained by applying various types of such multiomic integration, in particular the ones developed in WP2, to describe finding in molecular-based medicine with particular attention to drug development and therapy testing. The seminars related to then ongoing work on the usage of the same dynamic simulation tool to understand mechanotransduction phenomena, relating them to mechanical manipulations used as a therapeutic approach in RA (acupuncture). For the second objective, CNR and PICB worked together to on the integration of 13 datasets from heterogeneous biochemical origin to compile the molecular network on rheumatoid arthritis. The exchange was fostered on site and remotely, afterwards, thanks to the good level of familiarity between the team members, notably C. Nardini, Z. Xiaoyuan and P Tieri who spent 2 months each in each other’s institutes. The collaboartions continued beyond physical exchanges in order to coordinate the completion of two further scientific papers from the project. In addition, partners from the EU and China have proposed and acted as editors for the publication of a Research Issue of Frontiers (http://journal.frontiersin.org/ResearchTopic/2280) fully devoted to the integration of multi-omic data, one of the major aims of the whole project. This work in fact fulfils and expands beyond the partners of the project the aim of the KEPAMOD project, by establishing a wider community network devoted to the exchange of knowledge in processing and analysis of multi-omic data. The results of the project (Paolo Tieri*, XiaoYuan Zhou and Lisha Zhu Christine Nardini*, Multi-omic landscape of Rheumatoid Arthritis: re-evaluation of drug adverse effects, Frontiers in Systems Biology, 2014, | doi: 10.3389/fcell.2014.00059) have been disseminated in the framework of such special Research Topic in order to maximize its impact. Whilst scientific results have been achieved through publications and the release of freely available software, the project has had further positive impact on the scientists involved. Specifically, international experience has helped in the development of all involved, improving their language skills and thus the ability to disseminate results. A PhD student from China succeeded in securing a fully funded post-doc in Europe, as a result of her international experience and European scientists are able to better collaborate with international teams, opening the door for further funding. In addition, the collaborations have continued beyond the end of the project and partners from the EU and China continute to work together in the field of mutli-omic data, drawing on the knowledge succesffuly exchanged as part of the KEPAMOD project.

Data: CORDIS, © European Union