ANR PRC DeepIntegrOmics project: “End-to-end deep learning for precision medicine based on metagenomics and data integration”, 2021-2025

ANR PRC PROJECT DeepIntegrOmics: End-to-end deep learning for precision
precision medicine based on metagenomics and data integration (01/10/2021-30/09/2025)

Summary

DeepIntegrOmics addresses the development of routine (“point-of-care”) prognostic or diagnostic models based on metagenomics. It focuses on two research questions in the use of deep learning for precision medicine based on multiple omics data 1) achieving reliable prediction from raw metagenomic data and 2) improving the accuracy of classification and stratification by integrating other omics data. Two further applied goals are to propose new approaches for 1) the identification of omic biomarkers of CMD stages and 2) for patient stratification by interpreting these neural network architectures from a single phenotypic database. We are using data from 1844 patients from the European H2020 MetaCardis project, and assessing the prognostic value of stratification to predict CMD progression for 807 of the 1844 patients for whom we have characterized their evolution (clinical changes) over 10 years. These objectives support translational and precision medicine (i.e. patient classification and re-stratification) with a view to deploying these models for routine use in clinical centers.

• Project start date: 01/10/2021
• Project duration: 48 months
• IBISC project leader: Blaise HANCZAR (Full professor Univ. Évry, IBISC AROB@S team)
• IBISC funding: 144 kEuros
• Project leader: Jean-Daniel ZUCKER (DR IRD)
• IBISC institutional partners: UMMISCO IRD, INSERM-SU UMRS 1269, LAMSADE – DAUPHINE PSL
• ANR project web page