Project

Back to overview

Towards a better understanding of disease through advanced analysis of large-scale data

English title Towards a better understanding of disease through advanced analysis of large-scale data
Applicant Bergmann Sven
Number 176138
Funding scheme Project funding (Div. I-III)
Research institution Département de biologie computationnelle Faculté de biologie et de médecine Université de Lausanne
Institution of higher education University of Lausanne - LA
Main discipline Genetics
Start/End 01.03.2018 - 28.02.2022
Approved amount 683'003.00
Show all

All Disciplines (2)

Discipline
Genetics
Methods of Epidemiology and Preventive Medicine

Keywords (11)

interactions; epigenetics; GWAS; predictive medicine; system biology; large-scale data; integrative analysis; gene expression; modules; Genomics; metabolomics

Lay Summary (German)

Lead
Genomweite Assoziationsstudien (GWAS) suchen nach Verbindungen zwischen Genotypen (d. h. Varianten der DNA-Sequenz) und phänotypische Merkmalen (wie Körpergrösse, Krankheitsanfälligkeiten oder auch die Konzentrationen von Molekülen in bestimmten Geweben). Um solche Verbindungen herzustellen, führt man statistische Datenanalysen von sehr vielen Proben durch. Klinische Kohorten sind Sammlungen solcher Daten von Menschen, gewöhnlich aus der gleichen Bevölkerungsgruppe oder an der selben Krankheit leidend. Die bevölkerungsbezogene Cohorte Lausannoise (CoLaus) und das Schweizer Nierenprojekt über Gene in Hypertonie (SKIPOGH) sind Schweizer Kohorten, die sehr reichhaltige Daten, von Genotypen über molekulare Phänotypen bis hin zu medizinisch relevanten Merkmalen des Organismus, liefern.
Lay summary
Unser Forschungsvorschlag ist auf dem Gebiet der computergestützten Biologie angesiedelt und konzentriert sich auf die Entwicklung und Anwendung neuer integrativer Ansätze zur Analyse von großen Datenmengen ("Big Data"), wie Genexpression (welche Gene aktiv sind), Metabolom (kleine Moleküle) und epigenomische Daten (chemische Modifikationen der DNA und ihrer Stützstrukturen, die ihre Sequenz nicht verändern) mit dem Ziel klassische GWAS zu verbessern. Ziel ist es, neue Einblicke in die molekularen Wirkungsmechanismen hinter Genotyp-Phänotyp-Assoziationen zu erhalten.

Unsere vorgeschlagene Forschung ist in drei Ziele unterteilt: (1) Wir werden unseren Pathway Scoring Algorithmus (Pascal) erweitern, ein Rechenwerkzeug, das wir entwickelt haben, um zu beurteilen, ob bestimmte Gruppen von Genen, die in biologischen Signalwegen zusammenarbeiten, mit GWAS Signalen angereichert sind. Unsere Erweiterung ermöglicht die Priorisierung von Signalen aus genetischen Varianten, die in DNA-Regionen liegen, für die bereits Informationen vorliegen, die ihre funktionelle Relevanz anzeigen. (2) Wir werden neue Metabolomics-Daten unter Verwendung von Kernspinresonanz (NMR) Experimenten von Urin- und Serumproben erzeugen, die zu verschiedenen Zeitpunkten von CoLaus-Probanden entnommen wurden. Unser Ziel ist es, zu untersuchen, wie sich diese Messungen und ihre zeitliche Veränderung im Zusammenhang mit Krankheiten, sowie anderen molekulare Daten, die von denselben Probanden verfügbar sind, verhalten. (3) Abschließend wollen wir epigenetische Profile untersuchen, die für SKIPOGH-Proben zur Verfügung stehen, um molekulare Wirkmechanismen und mögliche Rollen als Biomarker im Zusammenhang mit Krankheitsphänotypen aufzuklären. Wir werden auch eine Sammlung von Netzhautbildern von SKIPOGH-Probanden untersuchen und nach Merkmalen dieser Bilder suchen, die durch Genotypen erklärt werden können, da sie auch die Struktur anderer, weniger sichtbarer Organe beeinflussen können.

Unsere Forschung wird computergestützte Methoden im Kontext realer Daten, die von Schweizer Kohorten zur Verfügung gestellt werden, voranbringen, mit Fokus auf der Schaffung greifbaren und potenziell biomedizinisch relevanten Ergebnissen sowie nachhaltiger Werkzeuge, die für die Analyse von großen Daten anderer Kohorten nützlich sein werden.
 


Direct link to Lay Summary Last update: 24.01.2018

Lay Summary (English)

Lead
Genome-wide association studies (GWAS) screen for links between genotypes (i.e. variants of the DNA sequence) and phenotypic traits (such as human height, disease susceptibility or the concentrations of molecules in certain tissues). To establish such links one performs statistical analyses on data from large sets of samples. Clinical cohorts are collections of such data from human subjects, usually from the same population or selected based on a common disease. Specifically, the population-based Cohorte Lausannoise (CoLaus) and the Swiss Kidney Project on Genes in Hypertension (SKIPOGH) are Swiss cohorts providing very rich data ranging from genotypes, over molecular phenotypes to medically relevant organismal traits.
Lay summary

Our research proposal is set in the field of Computational Biology, focusing on the development and application of new integrative approaches for the analysis of large-scale data (“big data”), such as gene expression (which genes are active), metabolomic (small molecules) and epigenomic data (chemical modifications of the DNA and its support structures that do not alter its sequence) with the goal to enhance classical GWAS. Specifically, our aim is to provide new insights into the molecular mechanisms of actions behind genotype-phenotype associations.

Our proposed research is divided into three goals: (1) We will extend our Pathway Scoring Algorithm (Pascal), a computational tool that we developed previously to evaluate whether certain sets of genes annotated to function together in biological pathways are enriched in GWAS signals from a given trait. Our extension will allow for prioritizing signal from genetic variants that reside in DNA regions for which there is already information indicating their functional relevance. (2) We will generate new metabolomics data using Nuclear Magnetic Resonance (NMR) profiling of urine and serum samples taken at different time points from CoLaus subjects. Our aim is to study how these measurements, and their change in time, behave in the context of disease as well as other molecular data available from the same subjects. (3) Finally, we plan to investigate epigenetic profiles available for SKIPOGH samples in the context of disease phenotypes in order to elucidate molecular mechanisms of action and potential roles as biomarkers. We will also study a collection of retina images from SKIPOGH subjects, searching for features of these images that can be explained by genotypes, as they may also affect the structure of other, less visible, organs.

Our research will advance computational methodologies in the context of real data made available by Swiss cohorts with focus on generating tangible and potentially biomedically relevant outputs, as well as sustainable tools that will be useful for the analysis of large-scale data from other cohorts.

 
Direct link to Lay Summary Last update: 24.01.2018

Responsible applicant and co-applicants

Employees

Associated projects

Number Title Start Funding scheme
152724 From modules to models III: Towards a better understanding of disease through advanced analysis of large-scale data 01.10.2014 Project funding (Div. I-III)

Abstract

The Cohorte Lausannoise (CoLaus) and the Swiss Kidney Project on Genes in Hypertension (SKIPOGH) are Swiss cohorts providing very rich data ranging from genotypes, over molecular phenotypes to medically relevant organismal traits. Great financial and human investments have been made to set up these and other cohorts, yet many aspects of the gathered data have been understudied, in part because the appropriate methodologies for their analysis are still poorly developed. Here we propose to advance and apply new integrative approaches for large-scale data, such as gene expression, metabolomic and epigenomic data, with the goal to enhance classical genome-wide association studies (GWAS) and elucidate mechanisms of actions behind genotype-phenotype associations. Specifically, our proposed research has the following three objectives:*Enhancing pathway analysis using GWAS data: We propose to extend our pathways analysis tool Pascal to include the option of prioritizing sets of Single Nucleotide Polymorphisms (SNPs) defined based on their annotation. This will open new possibilities for analyzing GWAS summary statistics at the pathway level in different biological contexts, such as coding vs non-coding or transcribed vs non-transcribed (in a given tissue). Using data from more than 200 GWAS we will investigate systematically which SNP prioritization strategies can help identifying disease pathways within Pascal. We will apply our insights for new analyses using data from CoLaus and other studies. We will make Pascal with SNP prioritization available to the community.*Integrative analysis of CoLaus molecular phenotypes: Our first aim is to generate new Nuclear Magnetic Resonance (NMR) data from 1000 CoLaus serum samples taken at the first follow-up to open up a longitudinal perspective onto this type of metabolomics data. Serum data from baseline and follow-up will be analyzed individually and in combination using our metabomatching software. We also propose the novel concept of “pseudo-compounds” as metabolomics meta-variables which jointly associate with a genotype or a molecular trait, like gene expression. This and other approaches will facilitate the integration of gene-expression with metabolomics data.*Integrative analysis of SKIPOGH data: We will investigate the structure of methylation profiles available for SKIPOGH samples in terms of their local correlation structure. We aim to guide the identification of functional units using genetic association. Such units will then be analyzed in the context of disease phenotypes in order to elucidate molecular mechanisms of action and potential roles as biomarkers. We will also investigate whether genotypic association may help to identify relevant features of SKIPOGH retina images. The work we propose builds on existing tools we have published and suggests broadening their functionality and extending their area of applications. Importantly, we develop our methodologies in the context of real data made available by Swiss cohorts with focus on generating tangible and potentially biomedically relevant outputs. We anticipate that our tools will be applicable to data from other efforts generating large-scale data for samples from genotyped subjects.
-