A methylation genetic test measures chemical marks attached to DNA rather than changes in the DNA letter sequence alone. DNA methylation helps regulate whether genes are active, and abnormal patterns can cause disease even when the underlying sequence is unchanged. Clinical methylation testing has several distinct uses. Targeted assays diagnose imprinting disorders such as Prader-Willi, Angelman, Beckwith-Wiedemann, and Silver-Russell syndromes. Genome-wide blood testing can identify disorder-specific “episignatures” associated with some developmental conditions and help clarify uncertain sequence variants. In cancer, methylation profiles can classify difficult tumors, especially central nervous system tumors, or measure a specific promoter marker with diagnostic, prognostic, or treatment relevance. These applications should not be treated as one interchangeable test. A normal targeted imprinting result does not exclude all genetic causes, and a tumor classifier result does not establish an inherited condition. Correct interpretation requires the tested tissue, loci or classifier, assay method, mosaicism sensitivity, copy-number information, and the clinical question for which the assay was validated.
- DNA methylation changes gene regulation without altering the underlying DNA sequence.
- Targeted imprinting tests detect parent-of-origin methylation patterns at specific chromosome regions.
- Methylation-sensitive MLPA can evaluate methylation and copy number in the same assay.
- Genome-wide episignatures can support diagnosis or variant interpretation in selected rare disorders.
- Tumor methylation classifiers compare a cancer profile with validated reference classes.
- Abnormal methylation may be primary or caused by a deletion, uniparental disomy, sequence variant, or tumor process.
Table of Contents
- DNA methylation and genomic imprinting
- Major clinical uses of methylation testing
- Laboratory methods and specimens
- Interpreting imprinting-disorder results
- Interpreting constitutional episignatures
- Interpreting cancer methylation results
- Negative, mosaic, and uncertain results
- Follow-up and counseling
DNA Methylation and Genomic Imprinting
DNA methylation most commonly involves addition of a methyl group to cytosine bases at CpG sites. The mark can influence how tightly DNA is packaged and whether nearby genes are transcribed. Methylation is part of normal development, tissue specialization, X-chromosome inactivation, aging, and genomic imprinting. It is not inherently abnormal or equivalent to a mutation.
Genomic imprinting is a parent-of-origin system in which one parental copy of a gene or region is normally active and the other is normally silenced. The maternal and paternal chromosomes carry different methylation patterns at imprinting control regions. If the active parental contribution is missing, duplicated incorrectly, or marked like the other parent, dosage of imprinted genes can become abnormal.
Several mechanisms can produce the same abnormal methylation pattern:
- a deletion removes the active parental chromosome segment;
- both chromosome copies come from one parent, called uniparental disomy;
- an imprinting-center defect establishes or maintains the wrong mark;
- a sequence variant disrupts a gene that controls imprinting;
- postzygotic error creates mosaic methylation disturbance in only some cells.
This explains why methylation testing can confirm an imprinting disorder yet not always reveal its molecular subtype. In Prader-Willi syndrome, for example, a maternal-only methylation pattern at 15q11.2-q13 establishes the diagnosis but does not by itself distinguish paternal deletion, maternal uniparental disomy, and an imprinting defect [1]. Follow-up testing is required because recurrence risk differs greatly among mechanisms.
Methylation can also become abnormal across the genome. Some pathogenic variants in chromatin-regulating genes produce reproducible blood DNA methylation patterns called episignatures. Cancer cells likewise accumulate characteristic methylation changes that reflect cell of origin and tumor evolution. Although all three applications measure epigenetic information, their reference samples, algorithms, and clinical claims are different.
Major Clinical Uses of Methylation Testing
Imprinting-disorder diagnosis uses targeted methylation analysis when a phenotype suggests a condition caused by abnormal parent-of-origin regulation. Common examples include Prader-Willi syndrome, Angelman syndrome, Beckwith-Wiedemann spectrum, Silver-Russell syndrome, Temple syndrome, Kagami-Ogata syndrome, and pseudohypoparathyroidism-related disorders. The test may be first line because it detects several molecular mechanisms that sequence analysis alone would miss.
At chromosome 15, Prader-Willi syndrome results from absence of the normally active paternal contribution in the critical region, whereas Angelman syndrome commonly reflects absence of maternal UBE3A activity in the brain. A methylation assay recognizes the parental pattern but does not detect every cause of Angelman syndrome, because some pathogenic UBE3A sequence variants leave methylation normal. Clinical sensitivity therefore differs even when the same locus is tested.
At 11p15, Beckwith-Wiedemann spectrum and Silver-Russell syndrome involve growth-regulatory imprinting domains. Methylation-sensitive testing often evaluates both methylation and copy number because deletions or duplications can alter the result. Mosaicism is common in these disorders, and blood may not contain the abnormal cell line at a detectable level. Testing affected tissue can improve sensitivity in selected cases.
Constitutional episignature testing uses genome-wide methylation arrays or sequencing to compare blood DNA with reference signatures for known neurodevelopmental and congenital syndromes. It can support a suspected diagnosis, help reclassify a variant of uncertain significance in an epigenetic regulator, or suggest a related disorder when sequencing is inconclusive. Clinical networks have developed reporting recommendations because classifier strength, signature overlap, and patient age or blood-cell composition influence interpretation [2].
Cancer testing includes several separate applications. A targeted assay may measure promoter methylation of a gene such as MGMT in glioma or methylation-related silencing of mismatch-repair genes. Genome-wide methylation profiling can classify tumors by comparing their pattern with reference entities. This has become especially influential in central nervous system tumor diagnostics, where histologically similar tumors can have distinct molecular classes [3].
Screening and liquid biopsy research may use methylation markers because tissue-specific patterns can help detect and localize cancer signals. Performance claims from one screening assay cannot be transferred to diagnostic tumor profiling or hereditary testing. Positive signals require a validated clinical pathway.
Laboratory Methods and Specimens
Methylation testing requires a method that distinguishes methylated from unmethylated DNA. The most common approaches use restriction enzymes, bisulfite conversion, methylation-sensitive amplification, arrays, or sequencing.
Methylation-specific PCR treats DNA so methylated and unmethylated sequences can be amplified differently. It is efficient for a defined locus and can identify an abnormal parent-of-origin pattern. Quantitative versions estimate the relative methylated fraction and may improve mosaicism assessment.
Methylation-sensitive multiplex ligation-dependent probe amplification, or MS-MLPA, combines methylation-sensitive enzyme digestion with multiplex probe amplification. It can assess methylation at selected sites and copy number at the same locus. This dual capability helps separate an epimutation from a deletion or duplication, although it does not identify every uniparental-disomy or sequence-level mechanism. The related MLPA test guide explains dosage analysis in more detail.
Bisulfite sequencing or pyrosequencing uses chemical conversion to distinguish methylated cytosines and can quantify methylation across selected CpG sites. Assays differ in the region analyzed and the normal reference range.
Genome-wide methylation arrays measure hundreds of thousands of CpG sites. Constitutional episignature and tumor-classifier algorithms compare a sample’s pattern with reference cohorts using statistical or machine-learning methods. The result may include a classifier score or calibrated probability, but the threshold and meaning are platform specific.
Long-read and whole-genome approaches can sometimes evaluate sequence, structural variation, and methylation on the same DNA molecules. These methods are developing rapidly, but clinical use requires validation for the specific disorder and specimen [4].
| Specimen | Common application | Key limitation |
|---|---|---|
| Peripheral blood | Imprinting tests and constitutional episignatures | Tissue-limited mosaicism may be missed; blood-cell composition affects genome-wide profiles |
| Buccal cells or skin fibroblasts | Follow-up for suspected mosaic constitutional disease | Collection and laboratory reference ranges must be validated |
| Chorionic villi or amniotic fluid | Selected prenatal imprinting testing | Placental biology, tissue-specific methylation, and mosaicism complicate interpretation |
| Formalin-fixed tumor tissue | Targeted cancer markers and tumor classification | Low tumor content, necrosis, and degraded DNA can lower confidence |
| Fresh or frozen tumor | Research or clinical classification | Availability may be limited; reference method still matters |
| Plasma or other fluids | Cancer-marker detection or monitoring | Low analyte abundance and uncertain tissue source can limit sensitivity |
For tumor testing, a pathologist selects an area with sufficient viable tumor. For blood episignatures, age, transfusion, stem-cell transplantation, and hematologic disease may influence the profile. The laboratory should state quality controls, assay coverage, reference class, and limitations.
Interpreting Imprinting-Disorder Results
An imprinting report typically states whether the methylation pattern is normal, abnormal, or inconclusive at a named locus. It may also report copy number. The first interpretive question is whether the pattern establishes the clinical diagnosis; the second is whether it identifies the mechanism.
A positive Prader-Willi methylation result usually shows only the maternal imprint at 15q11.2-q13. This confirms absence of the paternal expression pattern but leaves several possible causes. A copy-number loss supports a paternal deletion. If copy number is normal, SNP microarray or polymorphism testing can evaluate maternal uniparental disomy, followed by imprinting-center analysis when indicated.
An abnormal Angelman-related methylation pattern shows absence of the maternal imprint and can detect maternal deletion, paternal uniparental disomy, or an imprinting defect. A normal methylation result does not exclude Angelman syndrome caused by a pathogenic UBE3A sequence variant. Sequence analysis and deletion/duplication testing may therefore follow when the phenotype remains convincing.
At 11p15, the report may describe loss of methylation or gain of methylation at specific imprinting centers, along with copy-number findings. The exact pattern helps classify Beckwith-Wiedemann spectrum or Silver-Russell syndrome and can influence tumor-surveillance recommendations. Because abnormal cells may be unevenly distributed, a normal blood result may warrant testing another tissue when lateralized overgrowth, segmental findings, or classic features are present.
A result that shows abnormal methylation at several imprinted loci is called multilocus imprinting disturbance, or MLID. Its clinical meaning is still evolving. Consensus recommendations advise expert evaluation, caution in prenatal testing, and management based on methylation abnormalities at loci directly associated with established imprinting disorders rather than assuming every altered locus causes additional disease [5].
The recurrence risk cannot be read from the methylation pattern alone. Most isolated epimutations and uniparental-disomy events are sporadic, but some imprinting-center deletions or inherited sequence variants carry substantially higher recurrence. Parental testing and targeted molecular follow-up are therefore part of a complete diagnosis.
Interpreting Constitutional Episignatures
A constitutional episignature is a reproducible genome-wide blood methylation pattern associated with a particular genetic disorder or group of related disorders. Many arise from pathogenic variants in genes that write, erase, read, or organize epigenetic marks. The assay does not simply search for one methylated promoter; it compares a high-dimensional profile with reference signatures.
A positive result means the sample matches a validated signature above the laboratory’s threshold. It can support the pathogenicity of a sequence variant when the gene and signature are concordant. It can also provide diagnostic direction for a patient whose phenotype and sequencing results are ambiguous. The report may describe a strong match, moderate match, or class-specific score.
A positive episignature is not necessarily sufficient by itself to identify the exact DNA variant. Some related conditions have overlapping signatures, and different variants in one gene may produce different profiles. The result should be integrated with clinical findings and sequence or copy-number testing. When the signature suggests a diagnosis not previously considered, targeted molecular confirmation may be appropriate.
A negative result means no validated signature was detected among the classes included in the test. It does not exclude a genetic disorder, a disorder without a known episignature, a weak or age-dependent signature, mosaic disease, or a pathogenic variant whose molecular effect does not alter blood methylation enough to classify. Reference libraries expand over time, so later reinterpretation may be possible.
An inconclusive result can reflect a low classifier score, overlap between classes, technical quality, unusual blood-cell composition, or a profile outside the reference space. Reports should avoid converting a borderline machine-learning score into a definitive diagnosis. Published clinical experience supports structured reporting that states the tested classes, classifier confidence, limitations, and need for correlation [2].
Episignature testing is generally performed on blood. It should not be assumed that a tumor methylation classifier, prenatal sample, or another tissue will generate the same constitutional result. Tissue-specific methylation is a core biological feature, not merely a laboratory nuisance.
Interpreting Cancer Methylation Results
Cancer methylation tests may answer a narrow biomarker question or a broad classification question. The report should make that distinction explicit.
A targeted promoter test measures methylation in a defined regulatory region. In glioma, MGMT promoter methylation can provide predictive or prognostic information in appropriate contexts, but methods and cutoffs vary. A positive result does not mean every tumor cell has complete gene silencing, and a negative result does not prove normal expression. The result must be interpreted with tumor type, age, molecular classification, and treatment plan.
Mismatch-repair evaluation may include assessment of MLH1 promoter methylation when a tumor has loss of MLH1/PMS2 protein expression or microsatellite instability. Methylation can support a sporadic mechanism, but the hereditary-risk algorithm depends on cancer type, age, family history, somatic variants, and other tests. See the MSI test guide for the distinction between tumor repair deficiency and Lynch syndrome.
A tumor methylation classifier compares the genome-wide profile with reference tumor classes. The output may include a calibrated score, predicted class, and copy-number profile derived from array intensity. A high-confidence match can confirm or revise a diagnosis. A low-confidence result may mean the sample is poor quality, contains too little tumor, represents a class absent from the reference set, or lies biologically between established classes.
Classifier results should not override morphology automatically. An integrated diagnosis combines histology, immunohistochemistry, sequencing, copy number, clinical location, and methylation. Discordance should prompt pathology review, specimen verification, and consideration of another tumor sample. Studies of brain tumor classification show substantial clinical value while also demonstrating that algorithm choice and reference data influence performance [3].
Methylation-based cancer screening or tissue-of-origin testing has a different performance framework. A positive signal may indicate increased likelihood of cancer but usually does not provide a histologic diagnosis. Follow-up imaging and tissue confirmation remain necessary. Population prevalence strongly affects the probability that a positive result is truly cancer.
Negative, Mosaic, and Uncertain Results
A normal methylation result must be interpreted within the assay’s scope. A targeted test only evaluates selected loci. It may exclude the methylation abnormality typical of a disorder while leaving sequence variants, small copy-number changes, other imprinting disorders, and non-genetic diagnoses unresolved.
Mosaicism is a central limitation. An epigenetic error that occurs after fertilization may be present in only a fraction of cells. If the abnormal fraction in blood is below the assay’s detection limit, the result can be normal despite disease in another tissue. Quantitative methods can detect some mosaic patterns, but sensitivity is not unlimited. Testing buccal cells, skin fibroblasts, affected tissue, or another validated specimen may be considered.
Prenatal interpretation requires special caution. Methylation patterns can differ among placenta, chorionic villi, amniocytes, and postnatal blood. Confined placental mosaicism and culture effects can complicate results. Broad multilocus testing is generally not offered routinely in prenatal settings because reference ranges and clinical significance are less established [5].
For genome-wide classifiers, uncertain results can occur when the score is near the cutoff or the sample fails quality metrics. A “no match” result is not proof that methylation is normal; it means the profile did not match a validated class strongly enough. Reports should distinguish technical failure from a valid negative classification.
Cancer specimens add tumor purity and heterogeneity. Necrosis, inflammatory cells, and normal tissue can dilute the tumor profile. Treatment can select a subclone or change cellular composition. A classifier trained on untreated primary tumors may not perform identically on heavily treated metastases unless validated for that use.
Finally, methylation is dynamic. Age, tissue type, environment, and disease state influence patterns. Clinical assays focus on marks with demonstrated stability and diagnostic relevance, but incidental differences outside the validated claim should not be overinterpreted.
Follow-Up and Counseling
After an abnormal imprinting result, the next goal is to determine the mechanism. Follow-up may include SNP microarray, parental polymorphism studies, uniparental-disomy analysis, sequencing of an imprinting center or causal gene, and parental testing. This step refines recurrence risk and can influence management.
After a normal imprinting result with persistent clinical suspicion, ask whether the assay detects all known mechanisms and whether another tissue should be tested. Angelman syndrome evaluation may require UBE3A sequencing after normal methylation. Beckwith-Wiedemann or Silver-Russell evaluation may require mosaic-sensitive testing or alternative tissue. A clinical geneticist can prioritize the options.
For an episignature result, review whether the signature matches the gene and phenotype. A positive result may support reclassification of a sequence variant, but the laboratory that issued the original variant report may need to perform the formal update. A negative result should be interpreted against the list of signatures included at the time of testing.
For a cancer result, ask whether the test measured a single promoter or used a classifier, what cutoff was used, and how the result changes the integrated diagnosis or treatment plan. A pathologist should reconcile discordant morphology and molecular findings. Current disease-specific guidelines should determine whether the marker is predictive, prognostic, diagnostic, or investigational.
Genetic counseling is recommended for imprinting disorders because recurrence may range from very low to substantial depending on the mechanism. Counseling should also address mosaicism, variable expression, family testing, reproductive options, and condition-specific surveillance. Tumor methylation findings usually do not imply inheritance, but certain patterns may trigger separate germline evaluation.
Retain the complete report, including specimen, method, loci, methylation percentages or classifier score, copy-number findings, and limitations. “Methylation positive” is not a complete interpretation. The clinically useful conclusion is which methylation pattern was found, in what tissue, by which validated method, and what diagnosis or decision that pattern supports.
Methylation findings must also be interpreted with attention to tissue and cell composition. A constitutional imprinting abnormality may be detectable in blood, while mosaic disease can produce a weaker or tissue-limited pattern. Tumor methylation results can be influenced by tumor purity, necrosis, inflammatory cells, and normal tissue in the specimen. Age-related methylation and treatment effects may further complicate some assays. For these reasons, an abnormal profile should be matched to the assay’s validated specimen type and clinical purpose. When the pattern is unexpected, repeat testing, testing another tissue, copy-number analysis, uniparental disomy studies, or sequencing may be needed to identify the underlying mechanism and determine whether the finding is constitutional, mosaic, or tumor-specific.
References
- Prader-Willi Syndrome. 2024. GeneReviews clinical reference.
- Diagnostic utility and reporting recommendations for clinical DNA methylation episignature testing in genetically undiagnosed rare diseases. 2024. Clinical study and recommendations.
- Comparison of DNA methylation based classification models for brain tumors. 2024. Comparative study.
- Clinical evaluation of long-read sequencing-based genome-wide DNA methylation profiling. 2025. Clinical evaluation study.
- Multi-locus imprinting disturbance (MLID): interim joint statement. 2024. Consensus statement.
Disclaimer
This article is for general education and cannot interpret an individual imprinting, episignature, or cancer methylation report. Test methods and classifier thresholds are not interchangeable, and results should be reviewed with the relevant genetics, pathology, or oncology specialist. Prenatal and hereditary implications require condition-specific genetic counseling.
