Inspired by
@jwmares who was in turn inspired by
@patrickc, I ran my old 23andme genome through Claude Code and... also learned some useful things about myself!
I pasted Justin's same, simple prompt, and Claude output the five findings that matter most, ranked by how much they should change my behavior.
Both of my maternal grandparents had colon cancer, so I kinda suspected as much, but it's still interesting to see an elevated CRC risk flagged in my genomic data.
My findings:
1/ You are a poor metabolizer of a large, common drug class
You carry one *2 allele (rs1799853 CT) and one *3 allele (rs1057910 AC). Under the CPIC scoring system that's an activity score of 0.5 — poor metabolizer, the bottom tier. About 1–2% of people of European ancestry land here. You also carry one copy of the low-dose VKORC1 allele (rs9923231 CT).
This is not theoretical. CYP2C9 clears NSAIDs (ibuprofen, naproxen, celecoxib, meloxicam, piroxicam), warfarin, phenytoin, glipizide/glyburide, and losartan. Standard doses sit in your system substantially longer than they do in most people's — which is a plausible partial explanation if you've ever found ibuprofen unusually effective, or unusually rough on your stomach.
The combination of CYP2C9 poor metabolizer and a VKORC1 low-dose allele means that if you are ever prescribed warfarin, a normal starting dose would be genuinely dangerous. This belongs in your chart and on a card in your wallet.
2/ The favorable Alzheimer's genotype — and a reason to stop trusting LDL-C
rs429358 TT means no ε4 allele at all. rs7412 CT gives you one ε2. ε2/ε3 carries roughly 40% lower Alzheimer's risk than the common ε3/ε3 genotype — this is the best common APOE result available, and it's about 11% of people.
The lipid side is the actionable part. ε2 impairs remnant-particle clearance: it lowers LDL cholesterol while raising triglyceride-rich remnants. Your LDL-C will systematically understate your atherogenic particle burden. If you've been tracking a reassuring LDL number, it's the wrong number for you.
Ask for ApoB and non-HDL cholesterol. For an ε2 carrier those are meaningfully better measures of risk than LDL-C, and most standard panels don't include ApoB unless you request it.
3/ Non-secretor — and it explains your gut panel
rs601338 AA is the W143X nonsense variant on both copies. You don't secrete ABO blood group antigens into your mucus, saliva, or gut lining. Roughly 20% of Europeans are non-secretors.
The consequence that matters: those secreted antigens are a primary food source for Bifidobacterium. Non-secretors characteristically run low on Bifidobacteria and carry a structurally different gut community. If your Tiny Health report flagged low Bifidobacterium, this is very likely why — and chasing it with Bifido probiotics or HMO supplements tends to produce poor, non-durable colonization, because the mucosal niche those strains want isn't there.
Two upsides: you're resistant to the dominant GII.4 norovirus strains — genuinely, that's the winter cruise-ship one — and you have lower H. pylori colonization risk. One caution: non-secretor status is associated with higher measured serum B12, so a normal B12 reading is less reassuring for you than for most people. Pair it with methylmalonic acid.
4/ You carry one copy of the celiac-permissive haplotype
rs2187668 CT tags a single HLA-DQ2.5 haplotype. You're negative for DQ8 (rs7454108 TT). Around 25–30% of Europeans carry DQ2.5, and it's necessary but nowhere near sufficient — lifetime celiac risk for a heterozygote is roughly 1–3%. This is a door being unlocked, not opened.
The one rule this creates: if you ever develop unexplained GI symptoms, iron-deficiency anemia, or unexplained neuropathy, get celiac serology before you try cutting gluten. Eliminating gluten first makes the antibody test unreliable and you'd have to do a weeks-long gluten challenge to get a clean answer.
5/ Elevated risk of colorectal cancer
rs36053993 CT at chr1:45797228 is heterozygous for MUTYH G396D, the most common European MUTYH pathogenic variant. Your other MUTYH site is clean (Y179C, rs34612342 TT).
Being a single-copy carrier is modest news — roughly a 1.2–1.9× colorectal cancer risk depending on which study you read, not the polyposis syndrome that comes from carrying two copies.
---
Pretty much everyone should do this?