Ruibang Luo

This study introduces QUEST, a model that uses H&E histology to predict spatial protein expression, enabling virtual staining and analysis across biological scales.

LongPhase-TO reconstructs somatic haplotypes from tumor-only long-read sequencing, improving variant calling accuracy and haplotype block size.

Eva is a foundation model for tissue imaging data that learns multi-scale spatial representations and performs various tasks like cross-modal inference and patient stratification.

ClairS is a deep-learning method for long-read somatic variant calling that outperforms existing tools on synthetic and short-read datasets.

The paper re-architected a variant caller into a client-server system to improve agent-driven variant calling efficiency and reliability.

Clair3 v2 improves variant calling accuracy for Oxford Nanopore Technologies sequencing by using move table data and dwelling time, achieving higher F1-scores for SNPs and Indels.

HOPE learns tumor microenvironment signatures from paired H&E and spatial omics data to improve prognostic stratification using H&E images alone.

This paper presents near-complete, haplotype-resolved assemblies of a pancreatic cancer genome and normal tissue, identifying extensive somatic variants and epigenetic changes in previously excluded regions.

This study uses mitochondrial variants as barcodes to track clonal dynamics and spatial organization in human hematopoietic organoids.

Clair3-RNA is a deep learning variant caller for long-read RNA sequencing data that achieves high accuracy in SNP detection across different platforms.

LongPhase-S improves somatic variant accuracy by jointly estimating tumor purity and recalibrating variants using somatic haplotyping in long-read sequencing.

ClairS-TO is a deep-learning method for long-read tumor-only somatic variant calling that outperforms existing tools and works with both long-read and short-read data.

Clair-Mosaic detects mosaic small variants from long-read sequencing data using deep learning, achieving high sensitivity and accuracy in both paired and single-sample modes.

Gungnir is a DNA storage codec that corrects errors through hash matching and biochemical constraints, achieving full data recovery from 20% erroneous bases without redundancy or high-coverage sequencing.

OmicsNavigator translates spatial omics data into testable biological findings using a large language model for hypothesis validation and annotation.

This study evaluates genomic language models' ability to predict personal gene expression using paired genome-transcriptome data, finding limited success in capturing inter-person variability but improved performance for seen genes in new individuals.

AutoPM3 automates PM3 evidence extraction from scientific literature using LLMs, improving variant interpretation for rare diseases.

The paper introduces cuteSV2, a long-read-based method for accurate structural variant regenotyping, outperforming existing methods in benchmarking.

EPInformer predicts gene expression by integrating promoter-enhancer sequences, epigenomic signals, and chromatin contacts, outperforming existing models in cross-chromosome validation and CRISPR experiments.

This paper introduces Governome, a blockchain-based genomic data management system that enables owner-governed privacy, encryption, and verifiable data sharing for genomic analysis.

CellContrast reconstructs spatial relationships in single-cell RNA-seq data using contrastive learning, outperforming existing methods and enabling accurate spatial analysis.

ClusterV-Web provides a user-friendly web interface for analyzing Nanopore long-read data to profile HIV quasispecies and generate drug resistance reports.

This paper introduces Dip3D, a method to reconstruct high-resolution diploid 3D human genomes from Pore-C data, achieving better performance than existing methods.

This study improves variant calling by integrating Oxford Nanopore and Illumina data using Clair3-MP, showing benefits in difficult genomic regions.

The paper introduces the EN-TEx resource, which provides personal epigenomes and variant-impact models across multiple tissues, revealing allele-specific loci and their associations with GWAS loci.

This study uses ultra-low coverage whole genome sequencing on embryo samples to perform GWAS and identifies genetic loci associated with gestational age.

This paper introduces Clair3, a deep learning-based variant caller that combines pile-up and full-alignment methods, achieving faster performance and better accuracy, especially at lower coverage.

This study compares Oxford Nanopore adaptive sequencing and PCR amplicon sequencing for detecting Mycobacterium Tuberculosis in metagenomic samples and calls variants, with an open-source analysis framework.

Duet improves structural variant calling and phasing using low-coverage Oxford Nanopore sequencing data, outperforming existing tools at 8X coverage and maintaining strong performance at higher coverage.

This study identifies novel human DNA sequences using long reads and shows that assembly-free methods capture more novel sequences than assemblers.