Sam Freesun Friedman

This paper develops a machine learning framework to impute cross-platform proteomic data, improving replication of epidemiological findings and protein quality scoring.

The paper introduces Latentverse, an open-source toolkit for evaluating latent representations, providing reports and visualizations on properties like clustering and robustness.

This paper develops generative models that register and fuse multiple medical imaging modalities, revealing more genotype-phenotype associations through GWAS and PheWAS of learned embeddings.

The paper uses machine learning to analyze myocardial T1 time in UK Biobank data, identifying genetic loci and pathways associated with interstitial fibrosis and related cardiovascular diseases.

The paper uses machine learning to analyze drug testimonials and link them to neurobiological and psychological effects, revealing distinct patterns associated with different substances.

The paper uses deep learning and genome-wide association studies to identify genetic loci associated with right heart structure and function in 40,000 individuals.

This paper develops a cross-modal autoencoder framework to integrate cardiac MRI and ECG data, improving phenotype prediction and enabling MRI imputation from ECGs.

This paper analyzes 6,850 hallucinogenic experiences to link drug-induced consciousness changes to brain receptor distributions using natural language and gene expression data.

The paper uses machine learning to quantify liver fat from MRI data and identifies genetic variants associated with liver fat accumulation and related diseases.

This study uses deep learning to analyze aortic dimensions from MRI and identifies genetic loci associated with aortic diameter, highlighting SVIL as a key gene and developing a polygenic score for aneurysm risk.