12 December 2022 in Proceedings of the National Academy of Sciences
Correlations in gene expression are used to infer functional and regulatory relationships between genes. However, correlations are often calculated across different cell types or perturbations, causing genes with unrelated functions to be correlated. Here, we demonstrate that correlated modules can be better captured by measuring correlations of steady-state gene expression fluctuations in single cells. We report a high-precision single-cell RNA-seq method called MALBAC-DT to measure the correlation between any pair of genes in a homogenous cell population. Using this method, we were able to identify numerous cell-type specific and functionally enriched correlated gene modules. We confirmed through knockdown that a module enriched for p53 signaling predicted p53 regulatory targets more accurately than a consensus of ChIP-seq studies and that steady-state correlations were predictive of transcriptome-wide response patterns to perturbations. This approach provides a powerful way to advance our functional understanding of the genome.
26 September 2022 in Proceedings of the National Academy of Sciences
Recent advances in single-cell assay for transposase accessible chromatin using sequencing (scATAC-seq) and its coassays have transformed the field of single-cell epigenomics and transcriptomics. However, the low detection efficiency of current methods has limited our understanding of the true complexity of chromatin accessibility and its relationship with gene expression in single cells. Here, we report a high-sensitivity scATAC-seq method, termed multiplexed end-tagging amplification of transposase accessible chromatin (METATAC), which detects a large number of accessible sites per cell and is compatible with automation. Our high detectability and statistical framework allowed precise linking of enhancers to promoters without merging single cells. We systematically investigated allele-specific accessibility in the mouse cerebral cortex, revealing allele-specific accessibility of promotors of certain imprinted genes but biallelic accessibility of their enhancers. Finally, we combined METATAC with our high-sensitivity single-cell RNA sequencing (scRNA-seq) method, multiple annealing and looping based amplification cycles for digital transcriptomics (MALBAC-DT), to develop a joint ATAC–RNA assay, termed METATAC and MALBAC-DT coassay by sequencing (M2C-seq). M2C-seq achieved significant improvements for both ATAC and RNA compared with previous methods, with consistent performance across cell lines and early mouse embryos.
24 August 2026 in Journal of the American Statistical Association
OPOscar Hernan Madrid Padilla
Oscar Hernan Madrid Padilla
Transfer learning enhances model performance in a target population with limited samples by leveraging knowledge from related studies. While many works focus on improving predictive performance, challenges persist in statistical inference. Bayesian approaches naturally provide uncertainty quantification for parameter estimates; however, existing Bayesian transfer learning methods are typically limited to single-source scenarios or require individual-level data. We introduce TRansfer leArning via guideD horseshoE prioR (TRADER), a novel approach enabling multi-source transfer through pre-trained models in high-dimensional linear regression. TRADER shrinks target parameters toward an adaptively weighted average of source estimates, while remaining robust to differences in source scale and correlation. Theoretical investigation shows that TRADER achieves faster posterior contraction rates than standard continuous shrinkage priors when sources are well-aligned with the target, while preventing negative transfer from heterogeneous sources. The finite-sample marginal posterior behavior of TRADER is established. Extensive simulations show that TRADER achieves inference performance no worse than using target data alone, performs comparably to a competing frequentist method despite using only summary-level source data, and offers substantial computational advantages. Application to a high-dimensional genetic dataset further shows TRADER's effectiveness in inference under strong multicollinearity. Supplementary materials for this article, including a standardized reproducibility guide detailing all materials required to replicate the results, are available online.
5 November 2025 in Research Square
Fine mapping aims to identify causal genetic variants with nonzero phenotypic effects. Leveraging genome-wide association study (GWAS) data from diverse ancestries enhances fine-mapping accuracy and resolution by exploiting differences in linkage disequilibrium (LD) and increasing sample sizes. However, existing multi-ancestry fine-mapping methods rely on discrete priors and assume that all causal variants are shared across ancestries -- an assumption that may not hold in practice. Although MESuSiE accounts for both shared and ancestry-specific causal effects, it requires a priori specification of prior probabilities for causal variant sharing. Moreover, methods based on discrete priors are prone to sub-optimal convergence. To address these limitations, we introduce Multi-AnCestry Heritability INducEd Dirichlet decomposition (MACHINE), a flexible Bayesian framework that employs a continuous prior to model both shared and ancestry-specific effects without restrictive assumptions. Importantly, we propose an approach to control false discovery rate (FDR) for fine mapping with GWAS summary statistics and out-of-sample LD matrices, a challenge not addressed by existing multi-ancestry fine-mapping methods. We further improve fine-mapping performance by incorporating functional annotations of variants using generalized LD score regression (g-LDSC). Simulation studies across diverse genetic architectures demonstrate robustness and superior FDR control of MACHINE + g-LDSC compared to existing methods. In the real data analyses, we applied MACHINE + g-LDSC to four lipid traits and schizophrenia, identifying previously unknown causal variants and depicting their genetic architectures across ancestries.