Multi-Omics Integration: HBV Mutation Biomarker Panel

Two point mutations in the hepatitis B virus core promoter, A1762T/G1764A and G1896A, track with rougher liver disease, yet clinicians have had little serum-level signal to tell one chronic infection from another. A study in iScience ran 108 treatment-naive chronic hepatitis B patients through multi-omics integration, pairing data-independent acquisition serum proteomics with untargeted metabolomics. One pathway kept surfacing across every comparison: cholesterol metabolism.

What Multi-Omics Integration Found in Serum

The team depleted the 14 most abundant serum proteins, then quantified 1,223 proteins by MS-based proteomics and close to 2,000 metabolites by UHPLC-Q-Orbitrap mass spectrometry. Samples were pooled nine at a time per run to damp patient-to-patient variation, a pragmatic tradeoff that trims individual resolution but steadies the group-level signal. Against wild-type infection, the double-mutant group carried 79 differentially expressed proteins and 226 altered metabolites. When the proteomic and metabolite-profiling layers were mapped onto shared KEGG pathways, cholesterol metabolism stood out as the one node perturbed across all three mutant groups.

Key Findings

• Cholesterol metabolism was the common thread: integrated pathway analysis flagged it in the G1896A, A1762T/G1764A, and double-mutant groups alike, a convergence that single-layer studies tend to miss.

• Three lipid-handling proteins dropped: ANGPTL4, APOC2, and CETP were all downregulated in the more aggressive mutant infections relative to wild-type.

• The matching metabolites climbed: triglyceride, free cholesterol, and taurocholic acid rose in the same patients, consistent with stalled reverse cholesterol transport.

• A compact panel separated the dangerous variants: these six molecules reached ROC AUCs near 0.995 for the A1762T/G1764A and double-mutant groups, though they barely distinguished G1896A alone.

Figure 1. Quality-control overview of the data-independent acquisition serum proteomics. Panel A shows the protein intensity density curve and Panel B the matching boxplot. Panel C is a principal-component analysis of the four groups, Panel D reports replicate Pearson correlation, and Panel E breaks down protein coverage. Panel F tallies identified peptides, unique peptides, and quantified proteins. Adapted from Chen et al. (2026), iScience.

Cholesterol Metabolism as the Shared Signal

Inside the cholesterol module, the picture was internally consistent. The proteins that normally move lipids out of circulation fell, while the lipids themselves accumulated. ANGPTL4 and CETP both regulate lipoprotein lipase activity and reverse cholesterol transport, so their coordinated loss points to a host system that can no longer clear cholesterol efficiently during infection with these variants. The data-independent acquisition strategy that Bruderer and colleagues (2015) helped standardize is well suited to this kind of consistent, low-missing-value quantification across many serum runs.

From Pathway to a Diagnostic Panel

Discovery means little without an orthogonal check, so the authors measured the same proteins and metabolites by ELISA in an independent set of 108 patients. ANGPTL4 and CETP stayed significantly lower in the aggressive mutant groups, and free cholesterol stayed high, with AUC values above 0.99 for the double mutant. The honest caveat is that the single-center cohort is small and the near-perfect AUCs hint at overfitting, which the authors name directly.

Why Single-Sample Integration Matters Here

Reading proteins and metabolites from the same serum draw is what let the authors line up ANGPTL4 with free cholesterol inside one pathway. When both layers come off a single injection, which the Omni-MS workflow at Dalton is built around, the batch-effect and sample-volume tradeoffs that usually complicate multi-omics interpretation mostly fall away. That single-sample discipline is what makes a six-molecule panel like this one worth carrying into a larger validation study.

Frequently Asked Questions

What is multi-omics integration?

Multi-omics integration combines two or more molecular layers, such as proteomics and metabolomics, from the same samples and analyzes them together. The point is to catch relationships a single layer misses, like a protein and a metabolite shifting together in one pathway.

Can serum proteins predict hepatitis B mutation severity?

In this cohort, yes, for the more aggressive core-promoter variants. A panel of three proteins and three metabolites separated the A1762T/G1764A and double-mutant infections from wild-type with high accuracy. It did poorly on the G1896A mutation alone, so the signal is variant-specific.

What samples does multi-omics integration need for biomarker work?

Often just a single biofluid draw, such as serum or plasma, run on both a proteomics and a metabolomics platform. Pulling both readouts from one aliquot cuts volume demands and batch effects, which makes blood-based multi-omics integration practical for clinical biomarker programs working with limited samples.

Conclusion

The cholesterol link is believable here because two mass spectrometry platforms and an ELISA round all point the same direction. What isn't settled is whether the six-molecule panel survives outside this single center, given the modest cohort and the overfitting risk the authors raise. For a biomarker team, the practical takeaway is that viral genotype and host lipid handling belong in one assay design rather than two.

Citation

Chen, Y., Lin, P., Jin, J., Deng, M., & Wei, D. (2026). Cholesterol metabolism dysregulated by key HBV mutations revealed through multi-omics profiling. iScience, 29(5), 115844. https://doi.org/10.1016/j.isci.2026.115844

Note

This blog post summarizes findings from the above-cited research. Figures are adapted from the original publication. For full details, please refer to the source article.

By Seungjun Yeo, Co-Founder and CEO at Dalton Bioanalytics. Specializing in multi-omics mass spectrometry for drug discovery and biomarker research.