Francesca Ieva

This study expands the human plasma proteome analysis using an enhanced SomaScan platform, identifying new genetic associations and disease mechanisms beyond the classical secretome.

This study uses 3D diffusion autoencoders to derive latent cardiac MRI phenotypes, finds they share a genetic basis with cardiometabolic traits, and shows predictive power for disease risk.

SAFE-LD estimates linkage disequilibrium from GWAS summary statistics using pseudo-genotypes that preserve in-sample LD while protecting privacy.

This paper introduces a dual adversarial deconfounding autoencoder to remove batch effects from multi-center radiomics data, improving prognostic model performance for Hodgkin Lymphoma.

This paper introduces hiPRS, a novel polygenic risk score method that models high-order SNP interactions, demonstrating superior performance in simulation studies and real data applications.

This paper introduces a Deep Survival EWAS method to estimate risk profiles from pre-diagnostic DNA methylation data, showing improved identification of biologically relevant pathways in breast cancer.