CoChat

Multi-omics approaches to disease

Open access
Hasin, Yehudit, Seldin, Marcus, Lusis, Aldons · 2017 · Genome Biology

The DOI matches its Crossref record; 185 of the 186 references it cites were found in a registry or an index; 78 of the 108 sources its claims cite were checked for support.

Abstract

High-throughput technologies have revolutionized medical research. The advent of genotyping arrays enabled large-scale genome-wide association studies and methods for examining global transcript levels, which gave rise to the field of "integrative genetics". Other omics technologies, such as proteomics and metabolomics, are now often incorporated into the everyday methodology of biological researchers. In this review, we provide an overview of such omics technologies and focus on methods for their integration across multiple omics layers. As compared to studies of a single omics type, multi-omics offers the opportunity to understand the flow of information that underlies disease.

Overview

CoChat Summary

TL;DR

This review surveys omics technologies and argues that integrating multiple omics layers (“multi-omics”) can reveal the flow of information from genetic, environmental, or developmental causes to disease consequences better than single-omics studies. It organizes integrative approaches into genome-first, phenotype-first, and environment-first frameworks and discusses design, statistical, and translational challenges.

Key findings

  • Single-omics analyses are mostly correlative; multi-omics integration is motivated by observations that GWAS loci explain only a fraction of heritability, common diseases often involve regulatory variation, and the same variants can yield different outcomes depending on environment and genetic background.
  • Genome-first example: at the FTO obesity locus, chromatin maps across 127 cell types narrowed the region to an adipose-specific enhancer; IRX3 and IRX5 expression correlated with risk haplotype across 20 risk-allele and 18 non-risk-allele carriers, and CRISPR-Cas9 editing of rs1421085 implicated a single nucleotide variant disrupting ARID5B binding.
  • Phenotype-first examples: in Alzheimer’s disease, profiling seven epigenetic modifications found AD-associated genetic variants enriched in immune-related enhancers but not promoters or neuronal-function enhancers; in cardiac arrhythmia, immunoprecipitation of five long QT proteins identified 584 co-precipitated proteins, with 12 genes overlapping 35 GWAS loci.
  • Environment-first and population examples: 25 diets in over 800 mice showed macronutrient ratio effects on cardiometabolic health; about 100 inbred mouse strains were used to study high-fat/high-sucrose diet; 800 individuals were analyzed for personalized glycemic prediction; reference resources include METSIM (~10,000 Finnish men aged 45–65), MuTher (several hundred female twins), and UK Biobank (500,000 people).

Methods

  • Narrative review, not a primary study; it describes genomics, epigenomics, transcriptomics, proteomics, metabolomics, and microbiomics technologies and their application to human disease.
  • No primary data or sample collection; evidence comes from cited published studies and consortia, including Roadmap Epigenomics, GTEx, ENCODE, UK Biobank, MuTher, and METSIM.
  • Synthesis organizes multi-omics integration into genome-first, phenotype-first, and environment-first approaches and discusses correlation/co-mapping, mediation and conditional analysis, network modeling, ID conversion, and future directions.

Limitations

  • Paper-stated: human omics studies face confounding (e.g., diet, lifestyle), limited sample sizes/power, batch effects, lack of agreed gold-standard pipelines, and difficulty distinguishing causal from reactive changes; animal models have limitations including single genetic backgrounds and incomplete recapitulation of human complex disease.
  • Paper-stated: standardization across studies is difficult (e.g., RNA-Seq comparability depends on genome version, annotation, and quantification tools), and coordinated consortia are needed for multi-scale data generation, analysis, and translation.
  • Evident: as a review, it presents no new empirical results, does not perform a systematic search or meta-analysis, and its examples are selected to illustrate concepts rather than to quantify effects.

Model-written summary. Check it against the paper before relying on it.

Integrity signals

Registry

Crossref · 10.1186/s13059-017-1215-1

Cites

185 of 185 references verified0 citation errors · 0 not checkable

Statistics

Nothing to recompute.

Figures

Fig. 1 Multiple omics data types and approaches to disease research. Layers depict different types of omics data (Box 1). Omics data are collected on the entire pool of molecules, represented as circles. Except for the genome, all data layers reflect both genetic regulation and environment, which may affect each individual molecule to a different extent. The thin red arrows represent potential interactions or correlations detected between molecules in different layers—for example, the red transcript can be correlated to multiple proteins. Within layer interactions, although prevalent, are not depicted. Thicker arrows indicate different potential starting points or conceptual frameworks for consolidating multiple omics data to understand disease. The genome first approach implies that one starts from associated locus, while the phenotype first approach implies any other layer as the starting point. The environment first approach (not shown) examines environmental perturbations
Fig. 1 Multiple omics data types and approaches to disease research. Layers depict different types of omics data (Box 1). Omics data are collected on the entire pool of molecules, represented as circles. Except for the genome, all data layers reflect both genetic regulation and environment, which may affect each individual molecule to a different extent. The thin red arrows represent potential interactions or correlations detected between molecules in different layers—for example, the red transcript can be correlated to multiple proteins. Within layer interactions, although prevalent, are not depicted. Thicker arrows indicate different potential starting points or conceptual frameworks for consolidating multiple omics data to understand disease. The genome first approach implies that one starts from associated locus, while the phenotype first approach implies any other layer as the starting point. The environment first approach (not shown) examines environmental perturbationsCC BY · PubMed Central
Fig. 2 Usage of omics applications to prioritize GWAS variants. Locus zoom plot for a complex GWAS locus shows several candidate genes could be causal. Heatmap using various omics approaches for evidence supporting or refuting candidate causal genes. Beyond literature queries for candidates, various omics technologies and databases can be used to identify causal genes, including: searching for expression in relevant tissues [173–175], summary data-based Mendelian randomization (SMR) [176], mediation analysis [177], conditional analysis [23], correlation analyses, searching for overlapping pQTLs [178, 179], and/or implementing epigenetic data to narrow candidates (discussed for FTO locus [16])
Fig. 2 Usage of omics applications to prioritize GWAS variants. Locus zoom plot for a complex GWAS locus shows several candidate genes could be causal. Heatmap using various omics approaches for evidence supporting or refuting candidate causal genes. Beyond literature queries for candidates, various omics technologies and databases can be used to identify causal genes, including: searching for expression in relevant tissues [173–175], summary data-based Mendelian randomization (SMR) [176], mediation analysis [177], conditional analysis [23], correlation analyses, searching for overlapping pQTLs [178, 179], and/or implementing epigenetic data to narrow candidates (discussed for FTO locus [16])CC BY · PubMed Central
Fig. 3 Genome first approach at FTO GWAS locus. Claussnitzer et al [16] combined genomics, epigenomics, transcriptomics, and phylogenetic analysis to identify the functional element, the causative SNP, and the downstream genes mediating the genetic effect at the FTO locus in obesity. Circles represent genes in the locus and yellow circles represent genes implicated by the respective omics data. a Genomics: the FTO locus, containing several genes (circles), harbors the most significant obesity-associated haplotype in humans. SNPs that are in linkage disequilibrium with the risk allele are color coded—blue represents the non-risk (normal) haplotype and red the risk haplotype. b Epigenomics: publically available epigenomic maps and functional assays were used to narrow down the original associated region to 10 kb containing an adipose-specific enhancer. Chromatin capturing (Hi-C) was used to identify genes interacting with this enhancer. c Transcriptomics: this technique was used to identify which of the candidate genes are differentially expressed between the risk and normal haplotypes, identifying IRX3 and IRX5 as the likely downstream targets. In addition, conservation analysis suggested that rs1421085 (SNP that disrupts an ARID5B binding motif) is the causative SNP at the FTO locus. CRISPR-Cas9 editing of rs1421085 from background (TT) to risk allele (CC) was sufficient to explain the observed differences in expression of IRX3 and IRX5. d Functional mechanism: correlation and enrichment analysis were then used to identify potentially altered pathways that were then confirmed by in vitro and in vivo studies
Fig. 3 Genome first approach at FTO GWAS locus. Claussnitzer et al [16] combined genomics, epigenomics, transcriptomics, and phylogenetic analysis to identify the functional element, the causative SNP, and the downstream genes mediating the genetic effect at the FTO locus in obesity. Circles represent genes in the locus and yellow circles represent genes implicated by the respective omics data. a Genomics: the FTO locus, containing several genes (circles), harbors the most significant obesity-associated haplotype in humans. SNPs that are in linkage disequilibrium with the risk allele are color coded—blue represents the non-risk (normal) haplotype and red the risk haplotype. b Epigenomics: publically available epigenomic maps and functional assays were used to narrow down the original associated region to 10 kb containing an adipose-specific enhancer. Chromatin capturing (Hi-C) was used to identify genes interacting with this enhancer. c Transcriptomics: this technique was used to identify which of the candidate genes are differentially expressed between the risk and normal haplotypes, identifying IRX3 and IRX5 as the likely downstream targets. In addition, conservation analysis suggested that rs1421085 (SNP that disrupts an ARID5B binding motif) is the causative SNP at the FTO locus. CRISPR-Cas9 editing of rs1421085 from background (TT) to risk allele (CC) was sufficient to explain the observed differences in expression of IRX3 and IRX5. d Functional mechanism: correlation and enrichment analysis were then used to identify potentially altered pathways that were then confirmed by in vitro and in vivo studiesCC BY · PubMed Central
Fig. 4 The flow of biologic information from liver DNA methylation to liver transcripts, proteins, metabolites, and clinical traits. A panel of 90 different inbred strains of mice were examined for DNA methylation levels in liver using bisulfite sequencing. CpGs with hypervariable methylation were then tested for association with clinical traits such as a obesity and diabetes, b liver metabolite levels, c liver protein levels, and d liver transcript levels. Each dot is a significant association at the corresponding Bonferroni thresholds across CpGs with the clinical traits and metabolite, protein, and transcript levels in liver. The genomic positions of hypervariable CpGs are plotted on the x-axis and the positions of genes encoding the proteins or transcripts are plotted on the y-axis. The positions of clinical traits and metabolites on the y-axis are arbitrary. The diagonal line of dots observed to be associated with methylation in the protein and transcript data represent local eQTL and pQTL. The vertical lines represent “hotspots” where many proteins or transcripts are associated with CpG methylation at a particular locus. Figure taken with permission from [180], Elsevier
Fig. 4 The flow of biologic information from liver DNA methylation to liver transcripts, proteins, metabolites, and clinical traits. A panel of 90 different inbred strains of mice were examined for DNA methylation levels in liver using bisulfite sequencing. CpGs with hypervariable methylation were then tested for association with clinical traits such as a obesity and diabetes, b liver metabolite levels, c liver protein levels, and d liver transcript levels. Each dot is a significant association at the corresponding Bonferroni thresholds across CpGs with the clinical traits and metabolite, protein, and transcript levels in liver. The genomic positions of hypervariable CpGs are plotted on the x-axis and the positions of genes encoding the proteins or transcripts are plotted on the y-axis. The positions of clinical traits and metabolites on the y-axis are arbitrary. The diagonal line of dots observed to be associated with methylation in the protein and transcript data represent local eQTL and pQTL. The vertical lines represent “hotspots” where many proteins or transcripts are associated with CpG methylation at a particular locus. Figure taken with permission from [180], ElsevierCC BY · PubMed Central

Verification

Claims the paper makes

35 Valid · 11 Not valid · 16 Not checked · 33 Own results and interpretations

49 more claims · 25 with their cited sources checked

Create a free account to see all 95 claims and to verify the 30 references still open.

See all claims — free account
  • It may state two separate claims as one. 26
  • Probably not a substantial claim. 20
  • The quote only partly states this claim. 1

Create a free account to see these claims.

Claims other papers make about this one

Each is checked against this paper’s own text: does it say what the citing paper says it does?

Self-check

Percentages, sums and reported differences re-computed from the paper’s own numbers.

Each contradiction between two passages listed here was confirmed by an independent check.

1 of 1

Citations

Cites

Show all 185 references

Cited by

See all 2,953

Metadata

Authors
Hasin, Y., Seldin, M., & Lusis, A.
Affiliations
University of California, Los Angeles (US)
Corresponding author
Lusis, Aldons
Venue
Genome Biology
Published
2017-05-05
Volume
18
Issue
1
Pages
83-83
ISSN
  • 1465-6906 (ISSN-L)
  • 1474-760X
OpenAlex
W2611138580
Cited by
2,953
References
186
Citations by year
2026: 650 · 2025: 467 · 2024: 395 · 2023: 328 · 2022: 315 · 2021: 317 · 2020: 225 · 2019: 152 · 2018: 84 · 2017: 18
FWCI
12.26
Type
journal article
Language
en
Open access
Licence
Locations
  • Genome biology (publishedVersion, cc-by)
  • PubMed Central (publishedVersion, cc-by)
  • PubMed (publishedVersion)
  • DOAJ (DOAJ: Directory of Open Access Journals) (submittedVersion)
  • eScholarship (California Digital Library) (submittedVersion, cc-by)
Funders
National Institutes of Health (HL114437, HL28481)
Topics
  • Life Sciences › Biochemistry, Genetics and Molecular Biology › Molecular Biology › Bioinformatics and Genomic Networks
  • Life Sciences › Biochemistry, Genetics and Molecular Biology › Genetics › Genetic Associations and Epidemiology
  • Physical Sciences › Environmental Science › Health, Toxicology and Mutagenesis › Health, Environment, Cognitive Aging
Keywords
multi-omics integration, omics technologies, proteomics, metabolomics, transcriptomics, genomics, disease mechanisms
MeSH terms
Humans, Computational Biology, Genetic Predisposition to Disease, Genome-Wide Association Study, Datasets as Topic

About these checks

Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.

The claim verdicts this report counts are experimental: their accuracy is not yet measured.

A language model drew the claims from the paper on 2026-10-05, under version 3 of the claim definition (model deepseek-v4.1-flash, prompt extract-v3).

A second model labelled each claim on 2026-10-05 (model jev-1.13-20260917, definition revision gold-guideline-v0.2).

Cited sources are checked with a language model’s help (gpt-6-sol and qwen3.8-omni-flash).

Statistics and arithmetic are recomputed, not judged.

Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.

What counts as a claim ›How we measure confidence ›How we check this ›

Are you an author, or do you see a problem with this page? Request a review

How to cite

Cite CoChat for the checks on this page, and the paper for its own findings.

This page
CoChat. (2026). Citation check: Multi-omics approaches to disease. https://app.cochat.ai/papers/10.1186/s13059-017-1215-1
The paper
Hasin, Y., Seldin, M., & Lusis, A. (2017). Multi-omics approaches to disease. Genome Biology. https://doi.org/10.1186/s13059-017-1215-1

Data sources

Bibliographic data: Semantic Scholar, OpenAlex, Crossref.

Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).

Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.