Jove
Visualize
Contact Us
JoVE
x logofacebook logolinkedin logoyoutube logo
ABOUT JoVE
OverviewLeadershipBlogJoVE Help Center
AUTHORS
Publishing ProcessEditorial BoardScope & PoliciesPeer ReviewFAQSubmit
LIBRARIANS
TestimonialsSubscriptionsAccessResourcesLibrary Advisory BoardFAQ
RESEARCH
JoVE JournalMethods CollectionsJoVE Encyclopedia of ExperimentsArchive
EDUCATION
JoVE CoreJoVE BusinessJoVE Science EducationJoVE Lab ManualFaculty Resource CenterFaculty Site
Terms & Conditions of Use
Privacy Policy
Policies

Related Experiment Videos

Comparison of different cell type correction methods for genome-scale epigenetics studies.

Akhilesh Kaushal1, Hongmei Zhang2, Wilfried J J Karmaus1

  • 1Division of Epidemiology, Biostatistics, and Environmental Health, University of Memphis, Memphis, 38152, TN, USA.

BMC Bioinformatics
|April 16, 2017
PubMed
Summary

Related Concept Videos

You might also read

Related Articles

Articles linked to this work by shared authors, journal, and citation graph.

Sort by
Same author

The role of epigenetic age acceleration and social disadvantage in cardiometabolic health in southeast Louisiana women.

Clinical epigenetics·2026
Same author

DNA Methylation at Birth Showing Age-Specific Association with Atopy in Children: A Prospective Longitudinal Study.

Epigenomes·2026
Same author

Lung Dose and Pneumonitis Risk Following Five-Fraction Breast Radiotherapy.

Journal of medical imaging and radiation oncology·2026
Same author

Subfecundity, Infertility Treatment, and Child Neurodevelopment.

JAMA network open·2026
Same author

Post-traumatic stress disorder moderates the association between BrainAge acceleration and GrimAge acceleration.

Neurobiology of aging·2026
Same author

Epigenome-Wide DNA Methylation Levels During Pregnancy: Associations with Parity Across Diverse Populations.

Epigenomes·2026

This study compared eight cell-type correction methods for DNA methylation analysis. Surrogate variable analysis (SVA) is recommended for identifying informative CpGs, especially when reference data are unavailable.

Area of Science:

  • Genomics
  • Epigenetics
  • Bioinformatics

Background:

  • Whole blood DNA methylation studies are crucial for understanding environmental impacts and clinical outcomes.
  • Cellular heterogeneity in whole blood can confound these genome-wide association studies (GWAS).
  • Existing algorithms for cell-type correction exist, but their comparative performance with newer methods is unclear.

Purpose of the Study:

  • To compare the performance of eight different cell-type correction methods for DNA methylation data.
  • To determine which methods are most effective for adjusting for cellular heterogeneity in whole blood.
  • To provide recommendations for selecting appropriate methods based on study goals and data availability.

Main Methods:

  • Eight cell-type correction methods were evaluated: minfi, Houseman et al., RUV, FaST-LMM-EWASher, ReFACTor, RefFreeEWAS, RefFreeCellMix, and surrogate variables (SVA).
Keywords:
Cell-type compositionCpG sitesGenome-scale DNA methylationSurrogate variables

Related Experiment Videos

  • Methods were assessed using DNA methylation data from whole blood, examining associations with prenatal arsenic exposure and cancer status.
  • Performance was evaluated using homogeneous data with known cell compositions and simulated data with latent cell types.

    • Main Results:

    • The SVA-based method demonstrated high agreement with most other methods, except FaST-LMM-EWASher.
    • The minfi package provided more accurate cell-type estimations than the Houseman et al. method on homogeneous data.
    • Simulation studies showed SVA offered good sensitivity and specificity; RefFreeCellMix had high sensitivity but low specificity; FaST-LMM-EWASher had low sensitivity but high specificity.

    Conclusions:

    • Surrogate variable analysis (SVA) is recommended for identifying informative CpGs, particularly when reference data are absent.
    • The minfi package's method is advisable when suitable reference data are available for cell-type estimation.
    • For analyses not focused on cell proportion estimation or when reference data are lacking, SVA is the suggested approach.