BiGCaT Science Cafe
|28 Feb||16:00-17:00||UNS50/ 4.140a||Chris Evelo and Friederike Ehrhart||European Joint Programme on Rare Diseases (EJP-RD)||Abstract||Slides|
|18 Apr||16:00-17:00||UNS50/ 4.140a||Egon Willighagen and Serena Bonaretti|| EW: RiskGone and NanoSolveIT: two new H2020 projects
SB: Transparent research: Open-access data, reproducible workflows, and interactive publications
|02 May||16:00-17:00||UNS50/ 4.140a||Serena Bonaretti||Follow-up discussion on conducting transparent research||Abstract|
|16 May||16:00-17:00||UNS50/ 4.140a||Wouter van de Worp (PUL)||Identification of miRNAs in skeletal muscle associated with lung cancer cachexia||Abstract|
|06 June||16:00-17:00||UNS50/ 4.140a||Laurent Winckers (MaCSBio)||Deciphering the role of inflammation in diseases – a network biology approach||Abstract|
|13 June||16:00-17:00||UNS50/ 4.140a||Chaitra Sarathy (@ MaCSBio) and Denise Slenter|| CS: Genome scale metabolic modeling
DS: Combining forces for Recon and Wikipathways
|27 June||16:00-17:00||UNS50/ 4.140a||Klara Scupakova (@M4I) and Benjamin Baluff (@M4I)|
13 June 2019 – Chaitra Sarathy and Denise Slenter
Genome-scale metabolic models are important in understanding metabolism at the cellular level and play a key role in flux prediction and metabolic engineering projects. Recon, a stoichiometry based genome scale model for humans, provides a collection of genes, reactions and their corresponding metabolites as one comprehensive knowledge-base, based on literature and databases through manual curation. Combining genome scale metabolic models with omics data has been performed in several studies, however the focus on mathematical modeling meant that classical analyses which are part of omics data analysis pipelines (Pathway Enrichment and Network Analysis) have rarely been applied.
In order to improve the usefulness of these models for pathway and network analysis, several problems need to be addressed, e.g. inconsistencies in (or lack thereof) mappings between identifiers from different databases and incomplete stereochemistry of the involved metabolites. Nonetheless, Recon provides detailed information – such as cellular location, transport reactions and occurrence in tissue (types) – commonly missing from pathway knowledge bases.
During this Science Cafe, Chaitra and Denise will present part of their respective PhD projects, where they combine forces to overcome the issues raised above. First, Chaitra will explain how she visualizes omics data on “pathways” derived from genome scale models. This is followed by Denise, who will present their project idea for combining Recon3D (the latest genome scale metabolic model for Homo sapiens) with WikiPathways, aiming to aid both the bioinformatics and metabolic modeling communities and their respective projects.
6 June 2019 – Laurent Winckers
Title: Deciphering the role of inflammation in diseases – a network biology approach
The immune system is the defense mechanism of the human body and essential for our survival. While the immune system often succeeds swiftly in the defense against pathogens, at some occasions, it has to respond more extensively to succeed. The immune system contains a myriad of processes including inflammatory processes, which are triggered by recognition of pathogens and/or dead or injured tissue material. Inflammation is a crucial process, which is often an initial, fast response, eventually leading to restoring the functional and phenotypic homeostasis of the affected tissue. However, the slightest dysregulation of inflammatory processes or an excessive, long-lasting inflammatory response can have disastrous effects on the affected tissue or even the body as a whole and can cause or progress a diseased state. Little is known about the exact role of inflammation in many diseases and it is therefore of high importance to map and analyze the complex network of inflammatory processes to study the exact molecular mechanisms in diseases.
Inflammation genes were acquired from DisGeNET (Name: Inflammation, UMLS CUI: C0021368, GDA-score > 0.01) and were used to create a protein-protein interaction network with the stringApp in Cytoscape. Additionally, we identified inflammation-associated pathways from three major pathway databases (i.e. KEGG, WikiPathways and Reactome). Pathways were clustered based on gene overlap, similarities within the pathways and literature. Finally, an inflammation network of all process clusters and their associated genes was constructed in Cytoscape.
Gene expression data was gathered for four different diseases, i.e. breast-cancer, obesity, rheumatoid arthritis and dilated cardiomyopathy. The data was integrated and visualized in both the protein-protein interaction network and the inflammation network to study the inflammatory state of the different diseases. The complete computational workflow was automated in the R and documented in Jupyter notebooks.
The protein-protein interaction network showed more pro-inflammatory effect in breast cancer, obesity and rheumatoid arthritis. In dilated cardiomyopathy, the opposite held true. When investigating the larger, integrated inflammation network, breast cancer inclined a more anti-inflammatory state. Cluster analysis showed a mix of up- and down-regulated inflammation associated processes for breast cancer, which complies with literature that cancer could yield both a pro- or anti-inflammatory state. For obesity and rheumatoid arthritis, cluster analysis showed a pro-inflammatory state for almost all clusters which is accompanied with literature findings. As for dilated cardiomyopathy, cluster analysis showed a clear down-regulation of inflammatory processes which is in line with literature describing later stages of the disease.
Inflammation is an essential part of the immune system and is associated with many diseases. The created inflammation network provides a tool to analyze and integrate molecular data, and enables researchers to further study the inflammatory state and role of inflammation in a disease. Additionally, this yields the possibility to compare effects on inflammatory processes between different diseases. This might lead to new and better insights concerning inflammation in association with specific diseased states.
16 May 2019 – Wouter van de Worp
Title: Identification of miRNAs in skeletal muscle associated with lung cancer cachexia
INTRODUCTION: Cachexia, highly prevalent in patients with non-small cell lung cancer (NSCLC), impairs quality of life and is associated with reduced tolerance and responsiveness to cancer therapy and decreased survival. MiRNAs are small non-coding RNAs that play a central role in post-transcriptional gene regulation. Changes in intra-muscular levels of miRNAs have been implicated in muscle wasting conditions. Here we aimed to identify miRNAs that are differentially expressed in skeletal muscle of cachectic lung cancer patients to increase our understanding of cachexia and to allow us to probe their potential as therapeutic targets.
METHODS: Vastus lateralis muscle biopsies were collected of newly diagnosed treatment-naïve NSCLC patients with (n=15) and without cachexia (n=11), and healthy controls (n=22). MiRNA expression analysis was performed using a Taqman MicroRNA array. In silico network analysis was performed on all significant differentially expressed miRNAs. Differential expression of the top-ranked miRNAs was validated using RT-qPCR and subjected to univariate and multivariate cox proportional hazards analysis using overall survival (OS) and treatment-induced toxicity data obtained during the follow-up of this group of patient.
RESULTS: A total of 754 unique miRNAs were profiled and analyzed in a subset of muscle biopsies of NSCLC patients with cachexia (n=8) and age- and sex-matched healthy controls (n=8). We identified 28 significantly differentially expressed miRNAs, of which 5 miRNAs were upregulated and 23 downregulated. In silico miRNA target prediction analysis showed 158 functional gene targets, and pathway analysis identified 22 pathways related to the degenerative or regenerative processes of muscle tissue. Subsequently, the expression of six top-ranked miRNAs was measured in muscle biopsies of the entire patient group. Five miRNAs were detectable with qRT-PCR analysis, and their altered expression (expressed as fold change, FC) was confirmed in muscle of cachectic NSCLC patients compared to healthy control subjects: miR-424-5p (FC=4.5), miR-424-3p (FC=12), miR-450a-5p (FC=8.6), miR-144-5p (FC=0.59), and miR-451a (FC=0.57). In non-cachectic NSCLC patients, only miR-424-3p was significantly increased (FC=5.6) compared to control. Although the statistical support was barely insufficient to imply these miRNAs as individual predictors of OS or treatment-induced toxicity, when combined in multivariate analysis miR-450-5p and miR-451a resulted in a significant stratification between short and long-term survival.
CONCLUSION: We identified differentially expressed miRNAs putatively involved in lung cancer cachexia. These findings call for further studies to investigate the causality of these miRNAs in muscle atrophy and the mechanisms underlying their differential expression in lung cancer cachexia.
2 May 2019 – Dr. Serena Bonaretti
Title: Follow-up discussion on conducting transparent research
We will talk about aspects of data life cycle, data harvesting and the FAIR principle, and data usage (licensing).
18 April 2019 – Dr. Egon Willighagen
Title: RiskGone and NanoSolveIT: two new H2020 projects
Maastricht University is partner in two new Horizon 2020 projects, one around risk governance (RiskGONE, project no. 814425) and the other about predictive toxicology (NanoSolveIT, project no. 814572) of nanomaterials. Our own vision is to have risk governance fully transparent and explicitly linked to the underlying reasoning, choices, and, basically, all experiments. The missions of these two projects can be summarized as streamlining the process to research governance (RiskGONE) and the support decision making with cutting edge (computational) predictive toxicology (NanoSolveIT). This involves access to underlying data (experimental and computational), explaining why FAIR is a clear theme in both projects, as well as understanding how nanomaterials interact with living organisms, explaining where the systems biology comes in. This presentation will give an informal overview of the tasks that our BiGCaT team in these two projects, and how they link to our other projects.
18 April 2019 – Dr. Serena Bonaretti
Title: Transparent research: Open-access data, reproducible workflows, and interactive publications
Open science and reproducible research have progressively gained attention in the scientific community as the main concepts to practice transparent and rigorous research. They are fundamental to assess the value of scientific claims, build on previous work with reliability and efficiency, and strengthen collaborations to improve and expand robust scientific workflows. In this new context, papers, intended as the tool we use to communicate our findings to our peers, are evolving to become more interactive by including links to data and code repositories for reproducibility. In this talk, I will discuss definitions and principles related to transparent research, and I will provide some guidelines on how to deliver open-access data, reproducible workflows, and interactive publications.
28 Feb 2019 – Prof. Dr. Chris. Evelo / Dr. Friederike Ehrhart
Title: A cross-omics analysis work package in the European Joint Programme on Rare Diseases (EJP-RD)
The day after this talk, February 29, is #RareDiseaseDay (yes, a rare day indeed 😉 ). We thought this would be a nice occasion to tell you about the new EJP-RD, a large project that will support research in rare disease aiming to substantially improve diagnosis and treatment. EJP RD is a 5-year project, it has 85 partners in 33 different European and non-European countries and the total budget is about 110M Euro; part of that reserved for calls opened by the project itself. One of the main pillars of this project will develop and test a bioinformatics framework and infrastructure for rare diseases. It will use earlier ELIXIR experiences on FAIR data and BBMRI experiences with private patient data catalogues, and EJP RD was already selected as a new driver project for the Global Alliance for Genomics and Health (GA4GH). After all that data organisation we will apply bioinformatics interoperability approaches for integrative data evaluation. The work package on this latter cross-omics and data integration approach is lead by the Department for Bioinformatics at UM.
Next to tool development, this work will include the following steps:
- Create disease-specific pathways on a new WikiPathways rare disease portal (see http:///raredisease.wikipathways.org for the first version). This will be done based on database and literature evaluation and in collaboration with existing European expert networks (ERNs), quite a few of these are also active in our MUMC.
- Evaluate omics and other relevant data-availability on selected rare diseases and perform pathway (enrichment) analysis (likely already improving understanding of disease mechanisms. Omics data types already identified include genomics (largely exomes and single gene variants), transcriptomics and metabolomics.
- We will combine affected pathways into data and knowledge supported rare disease networks, evaluate these for things like active nodes and make them available on NDEX.
- Allow extension of these networks with relevant regulatory information (e.g. transcription factors and miRNAs) and where available evaluate data on such regulatory factors.
- Use the networks to evaluate drug targets and thus come up with ideas for drug repurposing with some special interest in orphan drugs (building on our IMI collaborations).
- Evaluate the network for intrinsic lifestyle factors (e.g. micronutrients present in or known to affect the networks) or processes known to be affected by exercise (building on our NuGO and other nutrition-related collaborations).
- Allow extension with external environmental factors like chemical exposure (toxicology) and evaluate overlap with so-called adverse outcome pathways (building on toxicology collaborations).
- Create complete workflows and make these available, including component containers and specific networks resulting from the analysis.
In this project, we will collaborate with various partners, The ones with central roles are: Franz Schaefer (WP co-lead, Heidelberg representing the ERN research group, Peter-Bram t Hoen (leads proof of principle task, RUMC Nijmegen, Marco Roos (link to FAIR work, LUMC Leiden), Anais Baudot (nutrient evaluation in networks, INSERM-AMU, Aix en Provence), Alberto Mantovani (toxicology evaluation, Istituto Superiore di Sanita (ISS), Rome), plus currently about 20 other participants. It will probably be clear that the project also links to most, if not all, activities that we currently have at BiGCaT. We expect that this project can thus also become a driver project for the department as a whole making the interlinks between various activities more tangible.