The objective of the COVIDOMICS AI project is to answer the following question: Why after the same COVID-19 infection (viral variant) within the same age group, some people only suffer from a “simple flu” while others will eventually be hospitalised in resuscitation service (and for some will die)? To answer this question, a multidisciplinary team of researchers and clinicians associated with the Laboratory of Molecular ImmunoRhumatology INSERM UMR_S 1109 (Dir. Prof. Seiamak BAHRAM) will analyse a cohort of patients infected with different viral variants (Delta, Omicron, and others to come). Two groups of patients will be compared: non-critical patients hospitalised in the so-called conventional sector and critical patients of resuscitation services. These patients will be studied at a very important level of clinical and molecular detail through a so-called “multiomic” approach. By applying a statistical analysis involving artificial intelligence and deep learning, it will be possible to identify therapeutic and diagnostic targets to differentiate critical forms from non-critical forms. In the long run, combined with the accumulated knowledge on variants, this research could also help anticipate the clinical impact of future variants, including on the post-infectious form of COVID-19, known as COVID-long.