Can My 23andMe or AncestryDNA Raw Data Show If I Have the GSTM1 or GSTT1 Null Deletion
23andMe and AncestryDNA raw files cannot reliably call a GSTM1 null or GSTT1 deletion, because null means the gene is missing and a SNP chip reads letters, not copies.
No, a 23andMe or AncestryDNA raw data file cannot reliably tell you whether you have the GSTM1 null or GSTT1 null deletion. "Null" means the entire gene is missing from both copies of the chromosome, and a genotyping chip reads single-letter variants, not how many copies of a stretch of DNA you have. A lab copy-number test can answer it, and so can a whole-genome file if someone reads the sequencing depth.
What "null" actually means
GSTM1 sits on chromosome 1 and GSTT1 on chromosome 22. Both make glutathione S-transferase enzymes, which attach glutathione to reactive molecules so the body can clear them. In a large share of people the whole gene has been cut out, a stretch of tens of thousands of DNA letters, by a recombination event between near-identical flanking sequences.
So this is not a typo in the gene. It is a missing gene. Someone with two deleted copies is "null" and makes none of that enzyme. Someone with one deleted copy makes some. The deletion is common: roughly half of people of European ancestry are GSTM1 null, and around one in five are GSTT1 null, with the frequencies differing a lot between populations. Outside Europe the picture shifts: GSTT1 null is far more common in East Asians, roughly half or more, depending on the study.
What 23andMe or AncestryDNA raw data can show about a GSTM1 null or GSTT1 deletion
A chip works by testing a fixed list of positions and reporting the two letters found at each. That model breaks when the DNA is not there:
- A deleted region may simply give a no-call, shown as
--, and a no-call looks identical to a probe that failed for any other reason. - If you carry one deleted copy, a SNP inside the region can read as homozygous (say
AA) when you have one A and nothing on the other chromosome. The file says two letters, the truth is one copy. - The raw file does not include the signal intensities that could in principle reveal a missing copy. You get calls, not the underlying measurements.
- Coverage of these two genes by probes is thin, and which probes exist changes between chip versions.
Raw files list rsIDs, chromosomes and positions, not gene names, so you cannot search the file for "GSTM1". To look, filter the chromosome and position columns on the GRCh37 ranges, which most consumer files use: GSTM1 is at about chr1:110,230,000 to 110,236,000, and GSTT1 at about chr22:24,376,000 to 24,384,000. Or paste single rsIDs into healthosx.com/dna/lookup. You may find no rows in those ranges, or a few rows with scattered -- calls. Whatever turns up, it still cannot settle null status: it does not prove you are null, and it does not prove you have the gene.
| Reads copy number | Separates 1 copy from 2 | |
|---|---|---|
| 23andMe or AncestryDNA raw file | No | No |
| Proxy SNP in a report | Guess | No |
| Multiplex PCR in a lab | Present or absent | No |
| qPCR, digital PCR or MLPA | Yes | Yes |
| Whole-genome VCF (SNPs/indels only) | No | No |
| 30x genome: BAM/CRAM depth or CNV calls | Yes | Yes |
Scroll the table sideways for the rest →
General capabilities. Specific products and chip versions vary.
Why proxy SNPs and single-rsID reports mislead
Some detox-gene reports fill the gap with a stand-in. They pick a nearby SNP that tends to travel with the deletion, or a single rsID, and report "likely null". Linkage between a SNP and a deletion is imperfect, it varies by ancestry, and it cannot separate one deleted copy from two. A confident "GSTM1: null" printed from a raw file is a guess dressed as a result.
The opposite mistake is just as common: seeing a normal-looking genotype on a nearby SNP and concluding you have the gene. You may have one copy, or none.
What can actually show it
A clinical or research lab has two main approaches. A multiplex PCR gives a present or absent answer, which identifies a null but cannot tell one copy from two. Quantitative methods, such as qPCR, digital PCR or MLPA, count copies, so they separate zero, one and two.
A whole-genome sequence can do it too, but only if someone reads depth, not genotypes. A standard 30x genome covers each position about thirty times on average. Where you have two copies, depth across the gene looks like the rest of the genome. With one copy it roughly halves. With none, there should be little or no depth across the deleted region.
Three cautions apply:
- 1.You need an alignment or CNV file. A small-variant VCF (SNPs and indels only) does not contain this. You need the BAM/CRAM, a depth track, or the provider's CNV/structural-variant calls. Some pipelines, such as DRAGEN and some consumer WGS deliveries, also emit CNV or SV calls that can include these deletions.
- 2.The reference build matters. GRCh38's primary chromosome 22 follows a haplotype that carries the deletion, so GSTT1 sits only on an alt contig (chr22_KI270879v1_alt). The reference is a mosaic of clones from several donors, not one person. A genome aligned to GRCh38 without alt-aware mapping can therefore show near-zero GSTT1 depth even in someone who has the gene. With alt-aware alignment the GSTT1 reads land on that alt contig, so that contig is where you check depth.
- 3.GSTM2 to GSTM5 look alike. GSTM1 has close relatives with very similar sequence, so reads can map ambiguously. A careful copy-number caller handles that, a quick glance at a genome browser may not.
Does a null result mean something is wrong
No. It is not a diagnosis and not a disease. The enzymes handle compounds such as combustion products and the isothiocyanates in cruciferous vegetables, and other glutathione transferases partly cover for them. Studies linking GSTM1 or GSTT1 null to cancer, lung function or drug toxicity exist, but results have been inconsistent, effects are small when they appear, and they depend on smoking, diet and workplace exposure. Whether null status should change anything you do is still argued about.
A missing gene cannot be read from a chip that only reads letters, so treat any "null" call from a raw file as unverified.
If you want a real answer
- Ask a doctor or genetics service whether a copy-number test makes sense for your reason for asking, for example before a specific drug or exposure question.
- A copy-number test is usually ordered through a clinician or a research lab, and in most countries it is not sold as a stand-alone consumer test. Availability and cost vary, so ask the lab or clinic.
- If you have a 30x genome, ask the provider for the alignment file or its CNV/structural-variant calls, and have the depth over GSTM1 and GSTT1 checked on the right reference. On GRCh38 that means the alt contig for GSTT1.
- Be wary of any report that prints a null call without saying how it was measured.

The download steps for 23andMe, on the page itself.

Ancestry across 15 regions and 157 reference populations.
The free tool at healthosx.com/dna reads a raw 23andMe, AncestryDNA, MyHeritage or whole-genome VCF file in your browser and reports traits, medication response and carrier status. GSTM1 and GSTT1 are not in its report.
New to raw DNA files? Start with the full guide: 23andMe Raw Data Interpretation Free in Your Browser
Written by
HealthOSX Research