Ramnik J. Xavier

The paper links CMV exposure-associated Type 1 immunity to improved long-term survival in sepsis without shock, but not in septic shock.

StriMap models TCR-peptide-HLA interactions using structural and sequence features, achieving state-of-the-art performance and identifying disease-associated peptides in ankylosing spondylitis.

The paper introduces CASTIE, a method for context-dependent eQTL mapping in single-cell data, identifying regulatory variants missed by static models and linking them to disease traits.

The paper identifies N-acyl putrescines, particularly NOP, as immunomodulatory metabolites produced by gut microbiota that suppress inflammation in IBD models.

The paper develops a deep learning model to predict peptide-HLA II interactions for post-translationally modified peptides, showing accurate predictions on unseen modifications.

The paper maps chemokine landscapes in solid tumors using spatial transcriptomics, identifying conserved and tumor-specific patterns and revealing perivascular CCR7 dendritic cells as key sources of lymphocyte-attracting chemokines.

TissueFormer is a foundation model that predicts spatial gene expression from histology images at single-cell resolution and scales to diagnostic tasks.

The paper develops an optical screening system to identify small molecules modulating α-synuclein aggregation in cells and finds harman reduces aggregate formation.

This study identifies 619 IBD risk signals, resolves 81 high-confidence variants, and prioritizes 664 candidate effector genes using GWAS, eQTL, and exome sequencing.

This study uses exome and genome sequencing to identify 68 genes directly associated with inflammatory bowel disease, providing clearer genetic insights and therapeutic targets.

This study creates a spatially-resolved molecular atlas of human skin and uses deep learning to identify 26 niches, revealing novel immune memory structures in inflammatory skin diseases.

This study identifies intelectin-2 as an antimicrobial lectin that crosslinks mucins and kills microbes by binding glycans in mucus.

This study identifies and characterizes the IsmB enzyme, which catalyzes a key step in coprostanol formation by the human gut microbiome, and links its presence to lower serum cholesterol levels.

This study shows neonatal BCG vaccination expands IL-17/IL-22-producing γδ T cells, mediates protection against TB, and induces a microbiome signature linked to gut-lung axis immunity.

CAZyLingua uses protein language models to annotate carbohydrate-active enzymes in metagenomic data, outperforming traditional methods and revealing new annotations in disease-related datasets.

The paper uses population-scale single-nucleus ATAC-seq and RNA-seq to identify regulatory variants and their disease associations in immune cells.

This study examines how gut bacteria genomes adapt to industrialization, finding convergent genomic responses in stress response, cell envelope remodeling, and central metabolism across different populations.

The paper investigates how industrialization affects microbial and physiological traits in the human gut across diverse populations, finding homogenized microbiomes, reduced diversity, and impaired disease prediction accuracy in less industrialized groups.

This study shows that better adherence to LE8 metrics improves gut microbiota diversity and cognitive performance in the Framingham Heart Study participants.

The paper identifies genetic variants affecting innate lymphoid cells in the gut using single-cell RNA-seq and validates their regulatory roles across cell types.

This study introduces STAG-seq, a method combining transcriptomics and genotyping to functionally analyze genetic variants in immune cells, revealing their impact on gene expression and disease mechanisms.

The paper identifies NXPE1 as an acetyltransferase that modifies mucin glycans in colonic mucus, showing its role in modulating the mucus barrier against commensal microbes in ulcerative colitis.

The paper uses spatial transcriptomics to map cell types and their spatial organization in healthy and fibrotic human liver, identifying changes in hepatocyte populations during fibrosis.

The paper introduces photo-stereoprobes for expanding the ligandable proteome and demonstrates their use in discovering bioactive small molecules.

SAIGE-QTL is a tool for single-cell eQTL analysis that directly maps eQTLs without pseudobulk aggregation and improves power for detecting rare variants.

The paper introduces a cost-efficient sequencing method called Blended Genome Exome (BGE) that combines low-coverage whole genome and high-coverage exome data, achieving >99% concordance with 30x WGS for variant imputation and rare coding variant calling.

The paper maps cell-type-specific eQTLs for HLA genes across multiple tissues and identifies dynamic regulation of HLA-DQ genes in immune cells.

This paper introduces a spatial sequencing method to study host-microbiome interactions in the mouse gut using deep learning for spatial niche detection.

This paper introduces MIST, a deep learning framework for metabolite structure annotation from mass spectra, outperforming existing methods and identifying new metabolites in a disease cohort.