Francesco Paolo Casale

This paper introduces SpaceDX, a method that jointly identifies differentially expressed genes and their spatial localization in multi-subject spatial transcriptomics data, improving detection compared to traditional approaches.

The paper benchmarks gene expression reconstruction from single-cell latent representations, finding autoencoders perform best at low dimensionality and foundation models depend on decoder architecture.

The paper introduces a phenome-wide genetic framework to identify mechanisms of social effects by analyzing genetic correlations between indirect genetic effects (IGE) and direct genetic effects (DGE) in mice.

REECAP uses contrastive learning to link retinal aging patterns with genetic loci associated with eye diseases, identifying 178 loci including 27 linked to age-related eye diseases.

VariantFormer predicts gene expression from diploid genomes, outperforming existing models and improving eQTL effect size estimation across ancestries and tissues.

LEOPARD completes missing views in multi-timepoint omics data by disentangling content and temporal representations, outperforming existing methods on benchmark datasets.

BayesRVAT improves rare variant association testing by integrating multiple annotations, increasing power and detecting novel gene-disease associations in real data.

The paper maps tissue-specific mtDNA heteroplasmy and its associations with age and mtDNA gene expression using bulk RNAseq data from GTEx.

FuncRVP improves rare variant polygenic risk scores by integrating functional gene embeddings, enhancing performance for traits with higher heritability.

HistoGWAS performs genome-wide association studies on histological traits using AI tools and variance component models, identifying four genome-wide significant loci in GTEx data.

This paper introduces Differentiable Mendelian Randomization (DMR) to predict disease risk using genetic data and risk factors without longitudinal data, validated on type 2 diabetes and Alzheimer's.

The paper develops AbExp, a model predicting aberrant gene expression across tissues, achieving 12% average precision by integrating variant effects and tissue-specific expression variability.

The paper introduces COAST, a rare variant association test for allelic series, which improves power over SKAT-O in detecting gene-trait associations in lipid and cell count traits using UK Biobank data.

This study identifies indirect genetic effects (IGE) from peers in mice using igeGWAS and confirms IGE genes through knockout experiments.

The paper characterizes proteome variation across 202 human iPSC lines, identifies 700 pQTLs, and links pQTLs to disease-associated GWAS variants.

The paper uses single-cell RNA-seq of differentiating iPS cells from 125 donors to identify expression quantitative trait loci influencing gene expression during endoderm differentiation.

This paper introduces StructLMM, a linear mixed model for analyzing multivariate gene-environment interactions, demonstrating its effectiveness in detecting GxE signals in BMI and eQTL data.

The paper identifies genetic-environment interactions in skeletal muscle gene expression using allele-specific expression and clinical traits as environmental proxies.

LiMMBo enables efficient joint genetic analysis of high-dimensional phenotypes using linear mixed models and bootstrapping, showing power advantages and scalability.

The paper introduces iSet, a method for detecting gene-context interactions using linear mixed models, and finds it identifies more interactions than alternative methods, including a gene-sex interaction for C-reactive protein.

The paper measures and compares transcriptional and epigenetic variability across immune cell types and finds neutrophils exhibit higher variability linked to immune pathways and environmental exposure.

This study analyzes 522 human iPSCs to identify genetic sources of phenotypic variation, finding 5-25% of variation in traits like differentiation capacity is due to individual differences.