G
Genetic Archaeology
Basics

How to Read Raw DNA Data: Decoding rsIDs, Strands, and Genotypes Safely

Published // November 14, 2026 // 8 min read

How to Read Raw DNA Data rsIDs and Genotypes

Anyone who has taken a consumer DNA test has spotted that tempting download button: “Download Raw Data.” Out pops a huge plain-text file stuffed with cryptic rows like rs1801133  1  11856378  AG. Understanding those four simple columns prevents the most widespread mistakes people make when checking their own genetics.

Anatomy of a Single Raw-DNA Line

Consumer DNA kits (SNP microarrays) do not sequence all three billion letters of your genome—instead, they probe roughly 600,000 to 700,000 high-information positions called SNPs (Single Nucleotide Polymorphisms; learn how lab genotyping works). Each row reports four fields:

Column Example Row Exact Meaning
rsID rs1801133 Unique global NCBI dbSNP identifier pinpointing this exact DNA address
Chromosome 1 Chromosome 1–22, X, Y, or MT (mitochondrial DNA)
Position 11856378 Base-pair coordinate mapped onto a reference genome build (typically GRCh37/hg19 or GRCh38)
Genotype AG or CT Your two measured DNA nucleotides (Adenine, Thymine, Cytosine, Guanine)—one from mom, one from dad

Three Classic Mistakes When Reading Raw Files

Trap 1: Plus Strand vs. Minus Strand Flips (Why C677T Says AG!)

DNA is a double-stranded helix built on strict Watson-Crick pairing: wherever the forward (“plus”) strand reads A, the opposite (“minus”) strand holds T; wherever one reads C, the partner reads G.

That simple geometry sparks endless panic with famous SNPs like MTHFR rs1801133 (C677T): medical papers report alleles as C or T (matching the gene’s reading direction on the minus strand), whereas 23andMe and Ancestry export genotypes strictly on the human reference forward (+) strand—displaying G (for C) and A (for T)! An “AG” in your raw file is completely normal C/T heterozygosity—not a bizarre rare mutation.

Trap 2: Microarray “False Positives” on Rare Clinical Mutations

While commercial SNP chips hit >99% accuracy on frequent variants (LCT, CYP1A2, HERC2), scientific studies in Genetics in Medicine proved consumer arrays exhibit surprisingly high false-positive rates on ultra-rare pathogenic mutations like BRCA1, CHEK2 c.1100delC, or MLH1 due to chip probe clustering quirks. Never act medically on unverified consumer raw data without accredited clinical confirmatory sequencing (see home-test pros & cons).

🔍 Quick Quality Checklist Before Trusting Any Third-Party DNA Tool

Verify whether the database states the genome build (GRCh37 vs. GRCh38), explicitly locks forward-strand orientation, distinguishes statistical associations (e.g. TCF7L2 odds) from monogenic diagnosis, and cites primary ClinVar/CPIC/OMIM classifications.

Trap 3: Over-Reading Single SNPs Outside Polygenic Context

A handful of traits obey clean single-locus switches—like earwax type (ABCC11), lactose tolerance (LCT), or primary blue/brown iris hue (HERC2). Complex outcomes like athletic capacity, stress resilience (COMT), or cardiometabolic health weave together hundreds of variants alongside lifestyle, sleep, and nutrition.

A Simple Workflow for Exploring Your File

  1. Check Orientation First: Look up any rsID on NCBI dbSNP to confirm whether published literature quotes the forward (+) or reverse (−) strand.
  2. Verify With Real Lab Biomarkers: Follow curiosity in wellness genes with simple blood measurements (Lp(a), HbA1c, fasting homocysteine, vitamin D, ferritin).
  3. Confirm Rare Clinical Alerts Medically: Any actionable hereditary cancer, cardiac, or clotting signal from a consumer upload warrants clinical confirmatory testing through your doctor or genetic counselor.

Frequently Asked Questions

What does ‘–‘ or ’00’ next to an rsID in my raw file mean?

That indicates a technical ‘no-call’ where optical intensity at that specific chip probe fell below quality thresholds. Seeing 1-3% scattered no-calls out of 650,000 rows is completely normal.

Why are C/G or A/T genotypes extra tricky to read?

These are called palindromic (ambiguous) SNPs because flipping strands produces identical letter pairs (C/G swaps with G/C). Reliable third-party uploaders use population allele frequencies carefully—or skip ambiguous positions—unless array build metadata explicitly confirms orientation.

Are raw text downloads formatted similarly across 23andMe, AncestryDNA, and MyHeritage?

Yes—all deliver tab- or comma-separated plain text with rsID, chromosome, coordinate position, and two-letter genotype, though exact covered rsIDs differ across chip versions (e.g., Illumina GSA v3 vs. v5).

Keep Reading

Related Research Guides

← Back to All Guides