MTHFR rs1801133 (C677T): What the Genotype Means and What to Measure

rs1801133 is a common missense variant in MTHFR: c.665C>T, p.Ala222Val, known historically as C677T. It makes the methylenetetrahydrofolate reductase enzyme thermolabile, which lowers its activity, which can raise plasma homocysteine when folate intake is low. About 10% of people of European ancestry carry two copies. For most of them the measurable consequence is a modest shift in one blood analyte that you can check directly for about $40. The genotype is not a diagnosis, and a genotype alone tells you very little that a homocysteine and folate panel would not tell you better. If you already have your sequencing data, the useful move is to confirm the call correctly, then go measure the biochemistry.

Getting the genotype right: strand, position, depth

The single most common error people make with rs1801133 is a strand flip. MTHFR sits on the minus strand of chromosome 1. The “C677T” naming is in transcript coordinates. On the reference genome, the alleles are G and A:

• GRCh38: chr1:11796321 G>A
• GRCh37/hg19: chr1:11856378 G>A
• HGVS: NM_005957.5:c.665C>T, NP_005948.3:p.Ala222Val

So the “TT” genotype appears in a GRCh38-aligned VCF as 1/1 with REF=G, ALT=A, and in a 23andMe or AncestryDNA raw text file as AA. GG is the wild-type homozygote. If you see a report claiming you are “CC” from a plus-strand array file, someone has translated coordinates by hand and you should re-check.

Pull it from a WGS VCF:

bcftools query -r chr1:11796321 \
  -f '%CHROM\t%POS\t%REF\t%ALT\t%FILTER[\t%GT\t%DP\t%GQ\t%AD]\n' \
  sample.g.vcf.gz

Three failure modes worth checking before you believe the result:

1. No line returned. In a variants-only VCF, absence means homozygous reference or means the site was never emitted. Those are different. Use the gVCF and confirm a reference block covers 11796321 with reasonable DP and GQ, or run samtools mpileup -r chr1:11796321-11796321 sample.cram and count reads yourself.
2. Low depth or unbalanced allelic depth. At 30x WGS you want DP of roughly 20 or more and AD near 50/50 for a het. An AD of 4,26 at a called het is a mapping artifact until proven otherwise.
3. Wrong build. If your pipeline emitted hg19 coordinates, 11796321 is a different base entirely. Check the VCF header ##reference line.

The neighboring variant people always ask about is rs1801131 (A1298C, NM_005957.5:c.1286A>C, p.Glu429Ala). It sits at chr1:11794407 T>G on GRCh38. It is about 1.9 kb from rs1801133 in the genome. If you are heterozygous at both, whether they are on the same chromosome or opposite chromosomes changes the interpretation: in trans, both copies of the enzyme carry a substitution. The 677T–1298C double-mutant haplotype in cis is rare in most populations, so compound heterozygotes are usually in trans, but you can check rather than assume. With short reads the two sites are too far apart to share a read pair, so use population phasing (shapeit4, or whatshap if you have any long reads or a trio).

What the substitution does to the protein

Ala222 sits in the catalytic domain near the FAD binding site. The valine substitution destabilizes the FAD interaction, so the enzyme loses cofactor more readily and is thermolabile in vitro. Reported residual activity is roughly 30% of wild type for TT and roughly 65% for CT under standard assay conditions. MTHFR converts 5,10-methylenetetrahydrofolate to 5-methyltetrahydrofolate, the methyl donor for remethylating homocysteine to methionine. Lower flux through that step means homocysteine accumulates and the folate pool shifts away from the 5-methyl form.

This is why the genotype effect is conditional on folate status. In folate-replete individuals the enzyme is stabilized enough that TT homozygotes have plasma homocysteine close to CC. In folate-deficient individuals the difference becomes large. Studies run in populations before mandatory folic acid fortification show much bigger genotype effects than studies run after. When you read an association paper on rs1801133, check the country and the year first.

Allele frequency varies substantially by ancestry. The T allele runs roughly 30 to 40% in Southern Europe and in Mexican populations, and around 10 to 15% in much of sub-Saharan Africa. It shows a clear geographic gradient across Eurasian populations, including marked variation among indigenous Siberian groups.¹ Any effect size you read from a cohort of one ancestry should not be assumed to transfer.

Measure the phenotype, which the genotype cannot give you

If rs1801133 matters for you, it matters through biochemistry you can measure. The panel we would run:

• Plasma total homocysteine, fasting. This is the direct readout of the pathway.
• Serum folate and red blood cell folate. RBC folate reflects the last three to four months, and serum folate reflects the last meal.
• Serum B12 and methylmalonic acid. Elevated homocysteine with normal MMA points at folate. Elevated homocysteine with elevated MMA points at B12.
• Creatinine, because renal function is a strong determinant of homocysteine and will confound everything else.

There is a real preanalytical failure mode with homocysteine. Red cells continue to export homocysteine into plasma after the draw, so a tube that sits at room temperature for an hour before centrifugation reads several percent high, and a tube that sits for several hours can read badly high. Ask the lab whether they separate plasma within 30 to 60 minutes or collect into a tube with an antiglycolytic additive. A “high homocysteine” result from a mail-in kit that spent two days in transit is a shipping artifact until repeated properly.

One caution about repeated testing: homocysteine has meaningful within-person variation, so a single value near a decision threshold is weak evidence. Two or three draws under matched fasting conditions is a better baseline. Interpreting any of these results, and deciding whether they warrant action, is a conversation with a clinician. A spreadsheet is not the place to work it out.

Where the association literature is solid and where it is thin

The homocysteine association is real and mechanistically clean. Downstream disease associations are weaker and heavily dependent on folate status in the population studied.

Cardiovascular and blood pressure associations have been reported repeatedly, including genetic association work linking MTHFR variants alongside WNK1 to hypertension susceptibility.² These are population-level effects, small per individual, and not a basis for treating a genotype.

The pharmacogenetic signal is the one that has the most direct bearing on care. MTHFR genotype has been associated with toxicity from high-dose methotrexate, an antifolate, in osteosarcoma patients.³ That is exactly the situation where genotype could be relevant to a treating oncologist, and it is the kind of result to carry to the physician managing the therapy rather than to act on yourself.

Associations with environmental sensitivity are also being studied, including work on MTHFR genotype distribution in populations with high dietary mercury exposure.¹ Treat that line of research as early.

What the literature does not support: the idea that rs1801133 requires a special diet, that it explains fatigue or mood symptoms in an individual, or that the genotype by itself calls for any specific supplement regimen. Claims of this kind are common on consumer genomics sites and are not supported by the association data.

Questions people also ask

What food should I avoid if I have MTHFR C677T? There is no evidence-supported list of foods to avoid based on this genotype. What the biology predicts is that the genotype effect on homocysteine is larger when folate intake is low, so the relevant question is total folate intake, which you can check with serum and RBC folate rather than by editing your diet on the basis of a SNP. Dietary changes intended to address a lab abnormality should be discussed with a clinician or dietitian.

What does MTHFR C677T cause? By itself, reduced MTHFR enzyme activity, and higher plasma homocysteine when folate is low. It does not cause a disease. Population studies link it to modest shifts in risk for several outcomes, including blood pressure phenotypes,² and to variation in response to antifolate chemotherapy.³

Is MTHFR something to worry about? For the great majority of carriers, no. Roughly 40 to 50% of people of European ancestry carry at least one T allele and about 10% carry two, which makes it a common variant rather than a rare finding. If you want a concrete answer for yourself, measure homocysteine, folate, B12, and creatinine, and take abnormal results to a physician.

Is a 23andMe call reliable for this site? rs1801133 is on the main genotyping arrays and calls well there, with the caveat that the output is plus-strand (G/A). Array calls at common, well-behaved SNPs are generally accurate, but confirming from WGS reads removes the strand and no-call ambiguity entirely.

Footnotes

1. B. A. Malyarchuk, N. V. Pokhilyuk. MTHFR Gene Polymorphism in Indigenous Populations of Siberia in Relation to Sensitivity to Mercury. Russian Journal of Genetics, 2026. https://doi.org/10.1134/s1022795426700481 ↩ ↩²

2. Laith Al-Eitan, Ola M. Al-Sanabra, Fouad A. Almomani, et al. Genetic Variants in the MTHFR and WNK1 Genes and Their Contribution to Hypertension Susceptibility. International Journal of Medical Sciences, 2026. https://doi.org/10.7150/ijms.133244 ↩ ↩²

3. Leilei Xu, Lujun Wang, Bingchuan Xue, et al. MTHFR variant is associated with high-dose methotrexate-induced toxicity in the Chinese osteosarcoma patients. Journal of bone oncology, 2018. https://doi.org/10.1016/j.jbo.2018.10.002 ↩ ↩²