α-Diversity analysis of hepatic transcriptome reveals distinct pathways in alcohol-associated hepatitis

Sudrishti Chaudhary, Jiajun Liu, Silvia S. Liu, Marissa Di, Juliane I. Beier, Rebeca Bataller, Josepmaria Argemi, Panayiotis V. Benos, Gavin E. ArteelJCI 2026 · DOI

Abstract

Next-generation sequencing can identify previously uncharacterized gene expression patterns in disease. Beyond differentially expressed gene (DEG) analysis, we investigated the ability of within-population diversity (α-diversity) of the transcriptome to reveal additional biological information in alcohol-associated liver disease (ALD), comparing differential Shannon diversity (DSD) to transcriptome heterogeneity changes. RNA sequencing data from normal livers and patients with early ALD and severe AH were analyzed. α-Diversity indices and percentage Shannon diversity of a gene, which refers to this gene's contribution to total Shannon entropy, were calculated. Ingenuity pathway analysis identified canonical pathways determined by DEG and DSD approaches. ALD significantly decreased hepatic transcriptome α-diversity, correlating with increased relative contribution of select genes. These changes were driven by lower-abundance gene expression loss. DEG and DSD analyses showed overlapping genes and canonical pathways, but DSD also identified additional genes and pathways not highlighted by DEGs, including fatty acid oxidation, extracellular matrix degradation, and cholesterol metabolism pathways that may represent additional therapeutic targets. Importantly, DSD more effectively identified differences between ASH and AH. Overall, α-diversity analysis revealed that ALD progressively reduces transcriptome heterogeneity, and that DSD provides complementary insights into disease mechanisms missed by standard approaches.

Introduction

Alcohol-associated hepatitis (AH) is a severe form of alcohol-associated liver disease (ALD) characterized by high mortality and limited therapeutic options. While RNA sequencing (RNA-seq) is a standard tool for investigating the molecular landscape of such diseases, traditional analysis typically focuses on Differentially Expressed Genes (DEG). DEG analysis identifies individual genes that are significantly up- or down-regulated but often overlooks the broader "ecosystem" of the transcriptome.

The authors of this study propose that chronic alcohol consumption acts as an environmental stressor that disrupts the liver's transcriptional complexity. By applying ecological biodiversity metrics—specifically α-diversity—to human RNA-seq data, they aimed to uncover biological mechanisms and therapeutic targets that remain hidden when using standard DEG approaches.

Methods and Study Design

The study utilized a multi-stage cohort and a dual analytical approach to compare healthy controls (HC) with patients suffering from alcohol-associated steatohepatitis (ASH) and severe alcohol-associated hepatitis (AH).

Workflow of the dual analytical approach comparing traditional DEG analysis with transcriptome α-diversity analysis.

Analytical Framework

The researchers compared traditional DEG analysis against a novel ecological framework:

  • Ecological α-Diversity Indices: Metrics such as the Shannon index, Menhinick richness, and Brillouin index were used to quantify the heterogeneity of RNA transcripts.
  • Differential Shannon Diversity (DSD): This core novelty involves calculating the Percentage Shannon Entropy (PSE) for each gene—its weighted contribution to total diversity. DSD is the log2 fold-change of these PSE values between conditions.
  • Bioinformatics Pipeline: Ingenuity Pathway Analysis (IPA) was employed to identify enriched biological pathways, and reference-based cellular deconvolution estimated shifts in liver cell populations.

Dataset

Secondary analysis was performed on publicly available RNA-seq data (dbGaP: phs001807.v1.p1) including 10 Healthy Controls, 11 ASH patients, and 18 AH patients.

Key Insights and Results

Transcriptome "Ecosystem" Collapse

The study found that a healthy liver maintains a highly diverse and evenly distributed transcriptome. As ALD progresses, this diversity collapses. Shannon diversity and Evenness indices showed a stepwise reduction from Healthy → ASH → AH, while the Dominance index peaked in severe AH. This suggests that a small subset of genes begins to monopolize transcriptional resources, much like an invasive species dominating a degraded ecosystem.

Vulnerability of Low-Abundance Genes

The loss of diversity is primarily driven by the silencing of low-abundance genes, likely due to epigenetic suppression or metabolic shutdown. Because DEG analysis often lacks the sensitivity to detect changes in low-expression genes, it misses a significant portion of the disease pathology that DSD successfully captures.

Unique Pathways Uncovered

While DEG and DSD analyses shared overlapping findings, DSD uniquely identified critical pathways in the AH vs. ASH comparison. These included:

  • Fatty acid oxidation
  • Extracellular matrix degradation
  • Cholesterol metabolism (highlighting regulators like MITF/p300/CBP and SPRING1)

Additionally, deconvolution analysis revealed a 19% loss of hepatocytes and an 84% increase in Kupffer cells/macrophages as the disease progressed.

Discussion and Future Directions

The application of α-diversity metrics reveals that ALD progressively reduces transcriptome heterogeneity. DSD acts as a powerful, complementary tool to DEG by detecting shifts in gene proportions even when absolute expression levels remain relatively stable.

Limitations: The study's cellular deconvolution relied on standard markers, which may not accurately reflect cells that have undergone disease-induced phenotypic changes (e.g., neutrophils in AH).

Future Work: Transcriptome diversity changes could serve as novel clinical biomarkers for ALD progression or treatment response. Further research is required to determine if this "diversity collapse" is specific to alcohol toxicity or a general response to severe cellular stress.

Summary updated

Abstract

Next-generation sequencing can identify previously uncharacterized gene expression patterns in disease. Beyond differentially expressed gene (DEG) analysis, we investigated the ability of within-population diversity (α-diversity) of the transcriptome to reveal additional biological information in alcohol-associated liver disease (ALD), comparing differential Shannon diversity (DSD) to transcriptome heterogeneity changes. RNA sequencing data from normal livers and patients with early ALD and severe AH were analyzed. α-Diversity indices and percentage Shannon diversity of a gene, which refers to this gene's contribution to total Shannon entropy, were calculated. Ingenuity pathway analysis identified canonical pathways determined by DEG and DSD approaches. ALD significantly decreased hepatic transcriptome α-diversity, correlating with increased relative contribution of select genes. These changes were driven by lower-abundance gene expression loss. DEG and DSD analyses showed overlapping genes and canonical pathways, but DSD also identified additional genes and pathways not highlighted by DEGs, including fatty acid oxidation, extracellular matrix degradation, and cholesterol metabolism pathways that may represent additional therapeutic targets. Importantly, DSD more effectively identified differences between ASH and AH. Overall, α-diversity analysis revealed that ALD progressively reduces transcriptome heterogeneity, and that DSD provides complementary insights into disease mechanisms missed by standard approaches.
BibTeX
@inproceedings{chaudhary2026diversity,
  author = {Sudrishti Chaudhary and Jiajun Liu and Silvia S. Liu and Marissa Di and Juliane I. Beier and Rebeca Bataller and Josepmaria Argemi and Panayiotis V. Benos and Gavin E. Arteel},
  title = {{α-Diversity analysis of hepatic transcriptome reveals distinct pathways in alcohol-associated hepatitis}},
  booktitle = {JCI},
  year = {2026},
  doi = {10.1172/jci.insight.200727},
}
Figure