How to Check MTHFR in Your 23andMe Raw Data, and Why It Is Not in GWAS-Based Reports

Your 23andMe raw file contains both common MTHFR variants: C677T is rs1801133 and A1298C is rs1801131. Search the file for each rsID and read the letters off the table below, remembering that 23andMe prints the plus strand, so C677T shows up as A, not T. Then read the evidence section: most health claims attached to MTHFR online are unsupported, and no GWAS-based report, ours included, scores these variants.

What MTHFR C677T and A1298C are#

MTHFR is a gene on chromosome 1 that codes for methylenetetrahydrofolate reductase, an enzyme that converts one form of folate into the form that circulates in the blood and is used to turn homocysteine into methionine (MedlinePlus Genetics, checked 14 October 2026). Two common variants change one amino acid each. C677T (rs1801133) produces an enzyme with reduced activity at higher temperatures, and people with two copies tend to have somewhat higher homocysteine. A1298C (rs1801131) has not been consistently linked to higher homocysteine on its own, according to 23andMe's review; it may matter when it sits alongside C677T.

Both are common polymorphisms, not rare mutations. The 1000 Genomes allele frequencies, read from Ensembl GRCh37 (rs1801133, rs1801131, checked 14 October 2026):

At those frequencies, roughly 13% of Europeans and one in five people of admixed American ancestry carry two copies of 677T, and about 60% of Europeans carry at least one. MedlinePlus puts two-copy carriers at "an estimated 25 percent of Hispanics and 10 to 15 percent of North American whites". "Having MTHFR" describes most of the population.

How to find them in your 23andMe file#

You do not need a website for this. Three steps:

1. Open the file in a text editor. Unzip the download (see our download guide) and open the .txt in TextEdit, Notepad or a code editor, not a spreadsheet.
2. Search for rs1801133. The line looks like rs1801133 1 11856378 AG: rsID, chromosome 1, position on build GRCh37, genotype. Note the two letters.
3. Search for rs1801131. The line looks like rs1801131 1 11854476 GT. Note the letters and read both off the table below.

The strand note matters here. The MTHFR gene is read from the reverse strand of chromosome 1 (Ensembl GRCh37 lists it on strand −1), but 23andMe reports every genotype on the plus strand of build 37; the header of your own raw file says so. The names C677T and A1298C use the gene's own letters, so in a 23andMe file the C of C677T is a G and the T is an A, and the A of A1298C is a T and the C is a G. SNPedia marks rs1801133 as a minus-orientation SNP for the same reason (checked 14 October 2026).

23andMe prints the two letters of a call in alphabetical order, so expect AG rather than GA, and -- means the array did not call that position. SNPedia's chip tags list rs1801133 on every 23andMe chip from v1 to v5 and on Ancestry's v2 chips. An AncestryDNA file splits the genotype into two columns; check its header for the strand statement before using the table. Our guide to reading a raw DNA data file covers the columns, builds and no-calls.

What the evidence actually shows#

Start with what is solid. C677T raises homocysteine modestly: 23andMe's literature review puts it at about 10% higher with one copy and about 30% with two, with most carriers still in the normal range (23andMe, "Our take on the MTHFR gene", updated 14 March 2025, checked 14 October 2026). The MTHFR Studies Collaborative Group put the TT-versus-CC difference at about 20% in the populations it studied (Clarke et al., PLoS Medicine 2012).

Then the part that matters: whether that translates into disease. The same 2012 study is the cleanest test. In 19 unpublished datasets totaling 48,175 coronary heart disease cases and 67,961 controls, the odds ratio for TT versus CC was 1.02 (95% CI 0.98–1.07). In the 86 published studies it was 1.15; the authors attributed the gap to publication bias and concluded that lifelong moderate homocysteine elevation has little or no effect on coronary heart disease.

Professional bodies drew the consequence a decade ago. The American College of Medical Genetics and Genomics published a practice guideline in 2013 stating that meta-analyses had found no association between MTHFR polymorphism status and venous thromboembolism, that the testing has minimal clinical utility, and that it should not be ordered as part of a routine thrombophilia evaluation (Hickey, Curry and Toriello, Genetics in Medicine 2013). In July 2015 ACMG put it on its Choosing Wisely list: "Do not order MTHFR genetic testing for the risk assessment of hereditary thrombophilia" (ACMG report, September 2015, hosted by NHGRI, checked 14 October 2026).

23andMe's own review reached the same place: the common variants "should not be interpreted as having a significant effect on health", because for most conditions studied the evidence is "inconclusive or conflicting". It names two exceptions. People with two copies of C677T appear to have a slightly higher chance of a pregnancy affected by a neural tube defect, with the overall chance still very low; and a handful of rare, different MTHFR variants cause homocystinuria, a serious disorder affecting fewer than 1 in 200,000 people. MedlinePlus agrees on the neural tube point and describes the other associations studied, from blood clots to psychiatric disorders and cancer, as having mixed results.

Why MTHFR is not in GWAS-based reports, including ours#

The GWAS Catalog is the public register of genome-wide association studies that our reports are built on. MTHFR is in it, but not for the things it is sold for. On 14 October 2026 the catalogue listed 48 associations for rs1801133, and most of them are blood measurements: homocysteine (7 rows, with p-values down to 10⁻¹²⁰), folate (5), red-cell size and hemoglobin content (12), blood pressure (4), glycine, reticulocyte volume, bilirubin. The rest are traits such as age at menopause, cortical thickness and kidney filtration rate. Only four rows map to a disease, each from a single study: multiple sclerosis (odds ratio 1.11, no confidence interval reported), early-onset schizophrenia (1.35, no interval), leprosy (0.81, no interval) and folate-deficiency anemia (an effect given as a beta in unnamed units). rs1801131 has one row in the entire catalogue, for hip circumference.

Our methodology keeps a catalogue row only if it is genome-wide significant, mapped to a disease, and reports an effect size with a plausible confidence interval. None of the four disease rows passes, and the measurement rows are not diseases, so neither rsID is in our health catalogue of 23,220 associations. (Our catalogues are built from the 2026-06-01 release; the folate-deficiency anemia row, a beta in unnamed units with no usable interval, would fail the same interval rule.) The nutrition catalogue does not contain MTHFR either: it tracks folate under the catalogue's "folate measurement" trait, the rs1801133 folate rows are filed under "folic acid level", and homocysteine is not a nutrient we track. Its seven folate-pathway rows (vitamin B6 and B12 levels) come from five other variants.

So, plainly: a Full Human report will not mention MTHFR, by design. It lists variants with a genome-wide significant, disease-mapped association and links each row to its study, as the sample report shows. If an MTHFR line is what you are after, the table above already gives it to you.

What Genetic Genie and other tools report instead#

• Genetic Genie runs a free, donation-funded methylation panel, the panel where MTHFR lives, and a detox panel. Your file is uploaded; the site says uploads are deleted within 24 hours and usually immediately after processing, and its privacy policy was last modified on 27 November 2019 (geneticgenie.org, privacy policy, checked 14 October 2026).
• Promethease ($25 per report, upload required; promethease.com, checked 14 October 2026) returns the SNPedia genotype page for rs1801133, which for AA carries a magnitude of 2.8 and lists studies pointing both ways (SNPedia, checked 14 October 2026).
• 23andMe itself now has a "Homocysteine (MTHFR-Related)" report covering C677T and A1298C for 23andMe+ Premium members, per the blog post above; its Health service is sold in eight countries.
• Xcode Life shows a free MTHFR status on upload and lists single reports from $30 to $69 (xcode.life, checked 14 October 2026); Genetic Lifehacks covers MTHFR in its gene articles, partly behind a $79.99-a-year membership, and says your DNA file is read by your own browser and never uploaded (geneticlifehacks.com, checked 14 October 2026).

Every one of these prints the same two letters you can read from the table, plus an interpretation, and the interpretation is where the evidence runs thin. Our comparison of raw DNA analysis tools covers price and where each tool processes your file.

What to do if you carry it#

Most people do, so start from that. If you are AG or AA at rs1801133, nothing in the guidelines above changes for you on the strength of the genotype alone; ACMG's position is that the test should not have been ordered for clot risk in the first place. If homocysteine concerns you, it is a blood test, not a genotype, and a doctor can order it and read it against your diet and B-vitamin status. Folate in food is the ordinary route. We do not give supplement advice, and neither should a website selling a "methylation protocol" based on two letters in a text file.

Two situations deserve a real conversation. If you are planning a pregnancy, the standard advice on folic acid does not depend on this genotype; discuss it with a doctor or midwife. If someone in your family has had homocystinuria or very high homocysteine, that points to the rare MTHFR variants, which an array file does not test and a clinical laboratory does. Either way, a doctor or genetic counsellor, not a raw file, is the next step.

FAQ#

Does 23andMe test for MTHFR?

Yes, in two ways. The raw file every customer can download contains rs1801133 (C677T) and rs1801131 (A1298C); SNPedia lists the first on every 23andMe chip from v1 to v5. Separately, 23andMe's "Homocysteine (MTHFR-Related)" report shows both variants for 23andMe+ Premium members, and its own review says the common variants should not be read as a significant health effect.

What is the rsID for MTHFR?

There are two common ones. C677T is rs1801133, at position 11,856,378 on chromosome 1 in build GRCh37; A1298C is rs1801131, at position 11,854,476. In a 23andMe file, AA at rs1801133 means two copies of C677T and GG at rs1801131 means two copies of A1298C, because the file reports the plus strand and the gene sits on the minus strand.

Is MTHFR a mutation?

The common variants are better called polymorphisms: 677T is carried by about 60% of Europeans and nearly three quarters of people of admixed American ancestry, with 13% to 22% carrying two copies. Rare MTHFR mutations that cause homocystinuria exist, affect fewer than 1 in 200,000 people, and are a different thing from C677T and A1298C.

What does AG at rs1801133 mean?

One copy of C677T. The G is the common version and the A is the 677T variant, written on the plus strand as 23andMe does. 23andMe's review puts homocysteine about 10% higher on average with one copy, usually still within the normal range, and ACMG's 2013 guideline finds no clinical use in testing for it as a clot risk factor.

Does Full Human report MTHFR?

No. Our reports use GWAS Catalog associations that are genome-wide significant, mapped to a disease and reported with an odds ratio or beta and a plausible confidence interval. rs1801133's catalogue rows are mostly blood measurements, and its four disease rows lack usable intervals, so neither MTHFR variant is in our catalogues. The report is built in your browser and nothing is uploaded, which you can verify yourself.