Premium
A network‐based, multi‐omics atlas for target identification and prioritization in Alzheimer’s disease
Author(s) -
Wörheide Maria,
Krumsiek Jan,
Nho Kwangsik,
Huynh Kevin,
Meikle Peter J.,
Saykin Andrew J.,
KaddurahDaouk Rima F.,
Kastenmüller Gabi,
Arnold Matthias
Publication year - 2020
Publication title -
alzheimer's and dementia
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 6.713
H-Index - 118
eISSN - 1552-5279
pISSN - 1552-5260
DOI - 10.1002/alz.045594
Subject(s) - omics , computational biology , computer science , biological network , data integration , interaction network , systems biology , profiling (computer programming) , disease , identification (biology) , data science , biology , bioinformatics , data mining , medicine , gene , genetics , botany , pathology , operating system
Background Alzheimer’s disease (AD) is an untreatable neurodegenerative disorder that represents a major global health burden. Large investments in the generation and analysis of increasing amounts of multi‐omics data have provided us with multi‐layered molecular imprints of pathogenic processes in AD. To identify the most promising targets for translating these readouts into development of novel therapies, it is crucial to take into account the complex biology underlying this heterogeneous disease via evaluating potential targets utilizing the entirety of multi‐omics evidences. However, integrated cross‐omics analyses are complicated and often lack intuitive and interpretable representation. Method We propose a network‐based, multi‐omics framework that allows for large‐scale analysis of data on biological entities across omics, as well as their associations with AD and related biomarker changes. Using Neo4j, a graph database, we store and interconnect these diverse biological domains in accessible network structures. Known biological relationships available in public databases such as gene‐transcript‐protein relations and pathway annotations constitute the backbone of this multi‐layered network. Large‐scale quantitative data from population‐based studies are included to establish data‐driven relationships within and across omics as a reference. To identify entities within this network that are relevant to AD pathomechanisms, the framework is finally overlaid with a comprehensive set of association data from multi‐omics screens in large AD cohorts generated through the AMP‐AD initiative. Result The resulting heterogeneous meta‐network is comprised of protein‐coding genes annotated with differential abundance in AD, readouts on more than 1000 metabolites and a large set of AD biomarkers and related phenotypes. Genes, proteins, metabolites, and AD phenotypes are interconnected by more than one million statistical links, including genetic and phenotype associations as well as gene/metabolite coexpression and co‐regulation data. We provide access to this comprehensive resource through an intuitive web‐interface, along with interactive visualizations and query and analysis tools of phenotype‐associated multi‐omics subnetworks. Conclusion The AD Atlas is a comprehensive and intuitive framework that facilitates the integration, exploration, and exploitation of large‐scale multi‐omics data. Applications of this resource include the network‐based identification of multi‐omics modules centrally linked to AD pathogenesis, followed by the prioritization of identified potential therapeutic targets for AD.