Eran Halperin

This paper introduces Unico, a method for deconvolving bulk genomic data into cell-type-resolved tensors, showing improved performance across gene expression and DNA methylation data.

The paper introduces ambimux, a method to estimate and demultiplex ambient molecule contamination in single-cell RNA+ATAC multiome data, improving the detection of biologically meaningful signals.

The paper characterizes the epigenomic and 3D genomic landscape of human subcutaneous adipose tissue at the cell-type level and links epigenomic variation to abdominal obesity risk.

This study performs a GWAS and evaluates polygenic risk scores for AMD in Israeli populations, finding consistent associations with known loci and demonstrating predictive power of international PRS.

The paper introduces CONTENT, a method for multi-context genetic modeling of transcriptional regulation, which improves power and discovers more gene-phenotype associations in TWAS.

The paper explores the relationship between genomic ancestry and health outcomes using a biobank linked to electronic health records, finding disease stratification across ancestry clusters.

This study evaluates methylation risk scores (MRS) in clinical datasets, showing they improve EHR imputation and predict phenotypes better than polygenic risk scores or EHR-based methods alone.

SNV-FEAST improves microbial source tracking by using signature SNVs instead of species abundance, showing better accuracy in three case studies.

This study shows that correcting background noise in microbiome data improves phenotype prediction by using unsupervised methods with appropriate transformations.

The paper benchmarks two methods for detecting cell-type-specific differential DNA methylation, finding TCA superior to CellDMC and discussing best practices for cell-type resolution analysis.

This study shows that low-coverage single-cell RNA sequencing of more samples increases statistical power for cell-type-specific eQTL analysis compared to high-coverage sequencing of fewer samples.

The paper introduces DIEM, a semi-supervised machine learning classifier to remove debris-contaminated droplets in snRNA-seq, improving clustering quality across three datasets.

ProPCA computes principal components for large genetic datasets efficiently, identifying novel signals of recent selection in the UK Biobank.

BATMAN integrates single-cell RNA-Seq datasets efficiently, outperforming existing methods by up to 80% in batch effect correction while preserving cell relationships.

Bisque estimates cell composition in bulk RNA-seq using single-cell data, improving accuracy and efficiency over existing methods.

This paper develops a machine learning model using EMR data to predict in-hospital mortality preoperatively, achieving higher accuracy than existing clinical scores.

The paper demonstrates a method to infer cell-type-specific DNA methylation from bulk data, revealing novel associations with leukocyte composition and rheumatoid arthritis.

The paper introduces CONFINED, a method to distinguish biological from technical variation in methylation data using multiple datasets, showing improved robustness and replicability.

This paper reviews statistical challenges in longitudinal microbiome studies and provides guidelines for design and analysis.

This paper introduces a Bayesian method to estimate cell type composition from DNA methylation without requiring a methylation reference, showing it can identify individual cell types rather than just linear combinations.

CRISPys is an algorithm for designing CRISPR sgRNAs that target multiple gene family members by leveraging the system's reduced specificity to simultaneously knockout homologous genes.

This paper develops a statistical method using genetic ancestry to detect gene-gene and gene-environment interactions in admixed populations, identifying significant interactions in methylation data.

RL-SKAT improves the accuracy of score tests for heritability and set-based genetic association by calculating exact p-values, avoiding the inflation or deflation of p-values seen in traditional methods.

Faucet constructs assembly graphs incrementally using a streaming algorithm, reducing disk usage and improving contiguity in metagenome assemblies.

This paper shows that DNA methylation data can capture ancestry information and introduces a method to infer ancestry from methylation data without genetic data.

This paper examines how ancestry-assortative mating affects linkage disequilibrium in admixed populations and shows it can lead to errors in demographic inference.

Recycler detects plasmids from de novo assembly graphs, improving recovery and accuracy over existing methods.

This paper introduces ALBI, a method for constructing accurate confidence intervals for heritability estimates using linear mixed models, addressing biases in existing approaches.