StudyRareStudyRare

Methylation Testing

Log in to star

Last updated Nov 10, 2026

Log in to add personal notes on this page.

Methylation testing measures whether cytosines at specific CpG sites carry a methyl group. Ordinary sequencing cannot see this mark, because PCR copies the base and leaves the methyl group behind. Clinical labs use methylation testing for three jobs: diagnosing imprinting disorders, completing fragile X testing, and reading the methylation of tumor genes that guides treatment. The biology of the mark itself is on Epigenetic mechanisms.

  • Bisulfite conversion: Sodium bisulfite turns unmethylated cytosine into uracil, which PCR then copies as thymine. Methylated cytosine resists the change and stays a cytosine. This turns a chemical mark into a sequence difference (C versus T) that a downstream method can read. Most of the methods below start here.
  • Methylation-sensitive restriction enzymes: Some enzymes cut their recognition site only when it is unmethylated. Comparing cut and uncut DNA shows the methylation state with no bisulfite step. MS-MLPA and the fragile X Southern blot work this way.
  • Methylation-specific PCR (MS-PCR): After bisulfite treatment, one primer pair matches the methylated sequence and another matches the unmethylated sequence. Which pair amplifies shows which alleles are present. It answers a yes or no question at one locus. See PCR.
  • Methylation-specific MLPA (MS-MLPA): An MLPA reaction split in two, with one half digested by a methylation-sensitive enzyme (HhaI). The undigested half gives copy number and the digested half gives methylation, so one assay shows both whether a region is deleted and whether its imprint is wrong.
  • Bisulfite sequencing: Sanger or NGS of bisulfite-treated DNA reads the methylation state of every CpG across a region.
  • Pyrosequencing: A sequencing-by-synthesis method that adds one nucleotide type at a time. Each base the polymerase adds releases pyrophosphate; ATP sulfurylase turns the pyrophosphate into ATP, and luciferase uses that ATP to make light. The light is proportional to the number of bases added, so the instrument reads the sequence as a series of light peaks. After bisulfite treatment, the ratio of C to T signal at each CpG gives the percent methylation at that site. Reads are short (typically under 100 bases), so it suits a defined set of CpGs rather than a whole gene. The same chemistry ran the first commercial NGS platform (Roche 454, retired in 2016).
  • Methylation arrays and episignatures: An array measures methylation at hundreds of thousands of CpGs across the genome. Many disorders of the chromatin machinery leave a reproducible genome-wide pattern in blood, called an episignature. A clinical episignature panel covers dozens of these disorders.
  • Long-read sequencing: PacBio and Oxford Nanopore read native DNA without PCR, so they detect 5-methylcytosine directly alongside the sequence. See Long-read sequencing.
  • Prader-Willi and Angelman syndromes: Methylation analysis at 15q11-q13 (SNRPN) is the first test, because deletion, uniparental disomy (UPD), and imprinting defects all produce the same abnormal pattern. It detects more than 99% of Prader-Willi syndrome and about 80% of Angelman syndrome. The Angelman cases it misses include UBE3A sequence variants, which leave methylation normal, so sequencing comes next.
  • Beckwith-Wiedemann and Russell-Silver syndromes: MS-MLPA at the two 11p15.5 imprinting centers is the first test for Beckwith-Wiedemann syndrome and Russell-Silver syndrome.
  • Fragile X syndrome: A full mutation silences FMR1 by methylating its promoter, so methylation status is part of the result. See Repeat expansion testing.
  • Finding the mechanism after an abnormal result: An abnormal methylation result says the imprint is wrong, not why. Copy number (built into MS-MLPA, or by microarray or FISH) finds a deletion. A SNP array or microsatellite studies with parental samples separate UPD from an imprinting defect. The mechanism matters because recurrence risk differs: a de novo deletion or UPD carries a low recurrence risk, while an imprinting center deletion inherited from a parent can carry a risk of up to 50%.
  • Tumor testing: MGMT promoter methylation in glioblastoma predicts benefit from temozolomide, an alkylating chemotherapy. The MGMT protein repairs the DNA damage temozolomide causes, so a tumor with a silenced gene cannot repair it. MLH1 promoter methylation in a colorectal or endometrial tumor that has lost MLH1 protein points to a sporadic cancer rather than Lynch syndrome. Pyrosequencing is a common method for both, and it also genotypes somatic hotspots such as KRAS codons 12 and 13 and BRAF V600E.
  • Variant classification: A variant of uncertain significance in a gene with a known episignature, such as KMT2D in Kabuki syndrome or NSD1 in Sotos syndrome, can be tested against that signature. A matching profile supports pathogenicity; a normal profile argues against it.
  • The tissue matters: Methylation is tissue specific, and a mosaic imprinting defect can be low level in blood. Beckwith-Wiedemann syndrome is often mosaic, so a normal blood result with a strong clinical picture leads to testing a second tissue, such as buccal cells or skin.
  • Methylation does not give the mechanism: Follow-up tests are needed to separate deletion, UPD, and imprinting defect, as above.
  • Bisulfite damages DNA: The treatment breaks much of the DNA, so the test needs more input than ordinary PCR.
  • Variants under primers or enzyme sites: A sequence variant in an MS-PCR primer site or in an HhaI site in MS-MLPA can mimic an abnormal methylation result.
  • Cutoffs are assay specific: Percent methylation is compared with each lab's validated thresholds, and values near a cutoff are reported as equivocal.
  • Methylation analysis catches Prader-Willi syndrome by every mechanism. Angelman syndrome has a gap: a normal result does not rule it out, because UBE3A variants leave the imprint intact.
  • Abnormal methylation with normal copy number means UPD or an imprinting defect, so the next test is a SNP array or parental studies.
  • An assay that reads light from a luciferase reaction is pyrosequencing, because every added base releases pyrophosphate that the enzymes turn into light.

"Pyro makes the light": pyrosequencing reads each base by the flash it produces, because every added base releases pyrophosphate, which ATP sulfurylase and luciferase turn into light.