Skip to main content

Validation of the MethylationEPIC BeadChip for fresh-frozen and formalin-fixed paraffin-embedded tumours.

Kling T, Wenger A, Beck S, Carén H

Publication: Clin Epigenetics (2017) • Vol 9 • Page 33PMID: 28392843View on PubMed

Purpose

This paper compared the correlations in methylation status between array types, frozen and formalin-fixed paraffin-embedded (FFPE) specimens and FFPE specimens subjected to different restoration methods. 

Conclusion

The average correlation in results between 450K arrays and those obtained using EPIC arrays was very strong for both the nine frozen specimens and the three FFPE specimens. While FFPE specimens were more likely to be of lower quality, the correlation between frozen and FFPE specimens was very strong. The specimens with the lowest correlations had a trend toward higher noise estimations. Restoration of FFPE specimens with the REPLI-g kit resulted in more middle range β-values, indicating a lower quality, but more probes failed with the Infinium kit and the average correlations with frozen specimens were comparable for the two methods.

Studies (2)

Purpose

This study investigated the effect of array type on methylation results, compared results obtained between frozen and FFPE specimens, and investigated indicators of lower correlations. Nine brain tumor specimens obtained during surgery were split and snap-frozen in liquid nitrogen or fixed in 4% formaldehyde for 24-48 h before dehydration and paraffin embedding. Three additional FFPE brain specimens were used to investigate the correlation between array types. DNA was extracted from frozen specimens using the DNeasy Blood and Tissue kit with an additional lysis step using the QIAGEN TissueLyser. DNA was extracted from FFPE sections (thickness not specified) using QIAamp DNA FFPE Tissue Kit with an additional deparaffinization in xylene followed by overnight digestion. DNA concentrations were determined by fluorometry and quality was assessed using the real-time PCR-based Illumina FFPE QC kit. FFPE specimens with a Δ cycle threshold of less than 3 were restored using the REPLI-g kit before bisulfate conversion using the EZ DNA methylation kit. Methylation was determined using Infinium Human Methylation 450 BeadChips (450 k) and Methylation EPIC BeadChips (EPIC). 

Summary of Findings

The average correlation in results between 450K arrays and those obtained using EPIC arrays was very strong for the nine frozen specimens (r=0.992, range r=0.988-0.996) and for the three FFPE specimens (r=0.984, range r=0.980-0.989). FFPE specimens were more likely to be of lower quality than frozen specimens but the correlation between frozen and FFPE specimens using EPIC arrays was strong (range from r=0.901 to r=0.987) with a mean of 0.938 and no effects of specimen age were observed. Importantly, the two tumors which showed lower correlations had lower levels of methylation in the FFPE specimen and there was a trend toward higher noise estimations in FFPE specimens with lower correlations to the matched frozen specimen. Further exploration of the 268 sites with a difference in β value of >0.2 for methylation in each of the nine frozen and FFPE specimens revealed that the differences were not localized to a particular area of the genome but were scattered.

  • Biospecimens

    TissueBrain
  • Preservation Types

    FormalinFrozen
  • Diagnoses

    NeoplasticNot specified
  • Platforms

    DNADNA microarray
    DNABisulfite conversion assay
  • Pre-analytical Factors

    ClassificationPre-Analytical FactorValue(s)
    Biospecimen PreservationType of fixation/preservation
    Formalin (buffered)
    Snap frozen
    StorageStorage duration
    14-45 months
    DNA microarray SpecificTechnology platform
    HumanMethylation450 BeadChip array
    Methylation EPIC BeadChip array