Leader : Dr Jean-Daniel Zucker
INTEGROMICS is the Bioinformatics Department of ICAN institute focused on integrating OMICS, Clinical and environmental data about patients to support the institute’s translational research mission. A strong emphasis is placed on the analysis of large scale biological datasets generated through the use of high-throughput technologies including microarrays, lipidomics, metabolomics and next-generation sequencing (genomics, transcriptomics, methylomics, metagenomics, …). It also includes radiomics, the integration and analysis of large amounts of advanced quantitative imaging features with high-throughput from medical images such as magnetic resonance imaging (MRI). INTEGROMICS also contributes to capitalizing and largely distributing existing and newly developed tools and methods at ICAN through a portal called IO::Portal. Finally, INTEGROMICS provides support to ICAN members and partners in the analyses of large scale data produced by different high-throughput OMICS technologies mentioned above. This structure was created in September 2015 to structure and rationalize the bioinformatics initiatives in ICAN and face up with the massive amount of data to be integrated and meant to support ICAN precision medicine.
INTEGROMICS research activities are structured around six main axes:
- DATA, PIPELINES and DATABASES: This axis covers data capture, curation, search, sharing, storage, protection, transfer, visualization, querying and privacy. It also concerns PIPELINES which transform raw data and ensure quality control but also the local maintenance of external public or private DATABASES (such as NCBI, KEGG, GO, etc.) used in these processes.
- HPC and SERVICES: This axis is concerned within ICAN with Solving complex computational problems associated to Big Data analysis. Parallel processing algorithms and intelligent heuristics are developed by incorporating both server administration and parallel computational techniques. It also covers the development of apps and tools that are shared with the scientific community.
- BIOINFORMATICS ANALYSIS: This axis concerns the development of various statistical analysis and bioinformatics pipelines dedicated to particular data types such as clinical, genomics, transcriptomics, metabolomics and metagenomics. Strong collaboration exists with ICAN Founders Advanced Research teams in bioinformatics.
- PREDICTIVE ANALYTICS: This axis is specifically concerned with advanced analytics to make predictions about a patient’s future condition or response to treatment and to identify multi-source signatures or medical scoring systems to support translating these predictions. The techniques used here are based on data mining, statistical modelling, high-dimension machine learning, and artificial intelligence. The produced models are aimed at being in fine translated to be used by clinicians and supports the treatment of the patient.
- NETWORK ANALYSIS and ANNOTATION: This axis revolves around the use of network abstractions to model, mine and integrate data from different sources. It also supports developing approaches to address the issue of vast amount of not or poorly annotated data. For instance, we use functional annotations from public databases (KEGG, GO) and propagate them onto gene co-expression networks in order to increase the accuracy of gene clusters and centrality measures to identify new biomarkers.
- SYSTEM’S BIOLOGY: This axis is concerned with the study of systems of biological components, which may be molecules, cells, organisms or entire microbial species viewed as a complex system i.e. whose behavior is difficult to predict from the understanding of its individual parts. Senior and expert biologist help us give biological insight and interpretation to our analysis results to devise new targets.
So far the INTEGROMICS department is composed of approximately 10 staff (1 senior researcher, 1 researcher, 1 associate professor, 4 postdocs, and 3 engineers) and 7 students (3 PhD students, 4 master students). A rapidly increasing number of tools and methods are being developed and will be shared with the community. The following list includes several applications developed or co-developed by Integromics members:
- FunNet: Integrative Functional Analysis of Transcriptional Networks. FunNet is an integrative platform for analyzing gene co-expression networks built from microarray expression data and biological annotations (GO, KEGG, etc.). It was developed in Jean-Daniel Zucker’s Bioinformatics team in Nutriomics initially by Corneliu Henegar as an R package and extended in 2008 by Edi Prifti with new functionalities such as annotation centrality measures, functional loads and a web service implementation. A Cytoscape plugin (FunNetViz) was also developed by Shakti Rielland that greatly improved FunNet result exploration.
- mQTL-NMR: Metabolomic Quantitative Trait Locus Mapping. mQTL-NMR is an R package developed and maintained by Lyamine Hedjazi. It provides a complete QTL analysis pipeline for metabolomic NMR data from preprocessing to metabolite identification. More details can be found at the mQTL-NMR package home page (click here).
- momr: Mining MetaOmics Data in R (MetaOMineR). ‘momr’ is the core package of a larger set of tools developed and maintained by Edi Prifti and Emmanuelle Le Chatelier for the analysis of quantitative metagenomics data. The package is structured around different modules such as preprocessing, analysis, visualization and contains routines for biomarker identification and pattern exploration.
- 3off2: A network reconstruction algorithm based on 2-point and 3-point information is an R Package developed by Séverine Affeldt at Isalembert’s Lab at Curie Institute. This algorithm combines constraint-based and Bayesian inference methods to reliably reconstruct graphical models, despite inherent sampling noise in finite datasets.
- predomics: Predictive Models for Omics Data. predomics is an R Package authored by Edi Prifti, Yann Chevaleyre and J.-D.Zucker based on three different heuristics to build scalable and efficient predictive model for Omics data and in particular for Metagenomics (to come).
- IO::Portal: ICAN IntegrOmics Interactive Portal. Developed by Tuong H. Nguyen the IO::Portal serves as a common ground for project management, data analyses and inhouse app building. Our apps are created using a modular technology and implemented as web services in the portal and are accessible by world’s scientific community.
INTEGROMICS will continue to develop new capabilities and anticipates becoming a center of expertise in the integrative bioinformatics and biostatistics analysis of high-throughput OMICS data. The department aims to achieve an international recognition in the field of cardiometabolics but also to become a key national player in complex human diseases, beyond the scope of focal IHU-ICAN topics. Links with other IHU (IHU-ICM) have already been established and will be strengthened in this regard.
To consult new positions for INTEGROMICS, click here.