Zihuai He

Unlocking Sensitive Data with SPHERE in the Age of AI

SPHERE enables the use of sensitive datasets in AI and open science by creating synthetic twins that preserve statistical properties and privacy.

Uncovering Heterogeneous Effects via Localized Feature Selection

This paper introduces a framework for identifying disease-associated features tailored to individual profiles using localized test statistics, applied to single-cell RNA sequencing data for Alzheimer's disease.

Beyond point estimates: quantifying predictive uncertainty reveals hidden dimensions of biological age acceleration and improves risk interpretation

The paper develops a framework to quantify predictive uncertainty in biological age estimates, revealing hidden dimensions of aging and improving risk interpretation.

Language Dysfunction Associated with Pathological Brain Connectivity in Children with Epilepsy

The paper identifies specific brain connectivity patterns in children with SeLECTS that predict language ability and distinguish epilepsy-related language dysfunction from normal variation.

Disease Outcomes in Boys with ABCD1 Variants Identified by Newborn Screening for X-ALD

The paper evaluates disease outcomes in boys with ABCD1 variants detected via newborn screening for ALD, finding that VUS variants are associated with lower rates of adrenal insufficiency and lower C26:0-LPC levels compared to pathogenic variants.

Plasma proteomics reveals divergent sex-specific senescence and bone biology signatures across neurodegenerative diseases

This study identifies sex-specific changes in bone and senescence-related proteins in neurodegenerative diseases using plasma proteomics.

Abnormal Elevated Connectivity During Language Processing is Associated with Poor Cognitive Performance in Children with Self-limited Epilepsy with Centrotemporal Spikes

The paper finds that children with SeLECTS show elevated motor-language connectivity during language tasks, which predicts worse phonological awareness.

Functional Connectivity in Self-limited Epilepsy with Centrotemporal Spikes (SeLECTS) Increases with Epilepsy Duration and Interictal Spike Exposure.

This study shows that functional connectivity in children with SeLECTS increases with epilepsy duration and interictal spike exposure, suggesting spikes drive neural network changes.

Divergent Electrophysiological Responses in the Human Hippocampus During Verbal Memory Processing

The paper examines electrophysiological responses in the human hippocampus during verbal memory tasks and identifies distinct functional clusters associated with memory encoding and retrieval.

Long-read genome sequencing and multi-omics in aging and neurodegeneration

The paper uses long-read genome sequencing and multi-omics data to identify structural variants associated with aging and neurodegenerative diseases, finding that SVs are more likely to be causal than SNVs.

Distilling Direct Effects via Conditional Differential Gene Expression Analysis

The paper introduces conditional differential gene expression analysis to distinguish direct from mediated gene effects in RNA sequencing data, finding most differentially expressed genes are not directly causal.

It's a wrap: deriving distinct discoveries with FDR control after a GWAS pipeline

This paper introduces solveblock, a software tool that enables FDR-controlled conditional independence analysis using GWAS summary statistics, leading to additional discoveries in British individuals.

Beyond guilty by association at scale: searching for causal variants on the basis of genome-wide summary statistics

This paper introduces a method to identify causal variants from genome-wide summary statistics, outperforming existing methods and discovering additional loci for complex traits.

Spinal cord injury: a systematic review and meta-analysis of microRNA alterations

This study performs a systematic review and meta-analysis of miRNA alterations in spinal cord injury across species, identifying conserved miRNAs involved in pathophysiological processes.

Single-soma transcriptomics of tangle-bearing neurons in Alzheimer's disease reveals the signatures of tau-associated synaptic dysfunction

The paper identifies gene expression signatures in Alzheimer's disease neurons with tau aggregates, revealing shared and subtype-specific molecular changes related to synaptic function and metabolism.

Multiple Causal Variants Underlie Genetic Associations in Humans

This study identifies multiple causal variants underlying genetic associations in humans using MPRA and eQTL data.

Summary statistics knockoff inference empowers identification of putative causal variants in genome-wide association studies

This paper introduces GhostKnockoff, a method to improve causal variant identification in GWAS by increasing statistical power and reducing proxy variants.

Powerful gene-based testing by integrating long-range chromatin interactions and knockoff genotypes

This paper introduces a gene-based testing framework that improves power and reduces false positives by integrating chromatin interactions and knockoff genotypes, showing better performance in GWAS and whole-genome sequencing data.

APOE missense variant R145C is associated with increased Alzheimer's disease risk in African ancestry individuals with the APOE ϵ3/ϵ4 genotype

Identification of putative causal loci in whole-genome sequencing data via knockoff statistics

This paper introduces a statistical method using knockoff statistics to identify causal loci in whole-genome sequencing data, showing increased discovery potential in Alzheimer's and COPD studies.

An evolutionarily acquired microRNA shapes development of mammalian cortical projections

The paper identifies a microRNA cluster unique to placental mammals that promotes corticospinal neuron identity by repressing LMO4, a callosal gene, thereby contributing to neuronal subtype diversity in the mammalian brain.

Multitrait genetic-phenotype associations to connect disease variants and biological mechanisms

The paper decomposes multitrait genetic associations to identify pleiotropy profiles linking SNPs to biological pathways and drug targets.

FUN-LDA: A latent Dirichlet allocation model for predicting tissue-specific functional effects of noncoding variation

FUN-LDA predicts tissue-specific functional effects of noncoding variants by integrating epigenetic data, showing higher accuracy than existing methods in eQTL and GWAS analyses.