11.2025 - 02.2026

PREDAMUS – PREDiction of lithium response using Algorythms based on MUltimodal Signatures

The PREDAMUS project aims to develop predictive models about the response to lithium in bipolar disorder (DB) with the integration of clinical and multi-omics elements (eg, genomics, epigenomics, transcriptomics). The study involves the retrospective evaluation of a total of 520 subjects and a prospective cohort of 80 subjects followed longitudinally and characterized in a standardized manner (ie, Alda Scale). Polygenic risk scores, epigenetic signatures as well as the content of neuronal extracellular vesicles enriched in microRNA will be calculated. The machine learning approach will allow exploring potential disease trajectories and respective molecular signatures associated with lithium response, potentially useful for identifying subjects characterized by a clinical phenotype with excellent lithium response, a fundamental prerequisite for optimizing precision psychiatry strategies.

If our efforts are successful, the results of our project can be translated into clinical practice to improve decision-making algorithms in prescribing lithium for bipolar disorder, if validation projects in independent cohorts confirm the data obtained. In this way, it would be possible hypothetically to reduce the time and risks associated with the trial-and-error approach that represents the therapeutic standard at the moment. Molecular signatures possibly identified could be the basis of tools to support clinical decision-making, promoting a personalized approach to pharmacotherapy in DB. The model could also potentially be extended to the prediction of response to other pharmacological interventions.

Settori Scientifico Disciplinari

BIO/14  PHARMACOLOGY MED/25  PSYCHIATRY

Spoke 5 : Mood and Psychosis