The letters you carry are not the trait you show
You asked about genotype vs phenotype because a file printed two letters and a website treated them as a fate. They are not a fate. They are a text. The body that lives in a world is the trait.
Think of a recipe versus a meal. The recipe is genotype: the alleles present at one or more sites. The meal is phenotype: what you can observe, measure, or diagnose, cooked in a kitchen you did not fully choose.
The real distinction is older than consumer DNA tests. Mendel needed it. Breeders needed it. You need it whenever a report confuses a variant with a person.
What genotype actually names
Genotype is the genetic constitution, usually at a specified place. At one SNP it might be A/A, A/G, or G/G. Across a gene it might be two alleles named by protein change. Across a genome it is the full set of such facts, which no everyday file actually contains.
Homozygous means both copies match at that place. Heterozygous means they differ. Those words are genotype words. They do not say whether a trait appears.
A consumer chip genotype is a sample of sites. A missing row is not a genotype. Write not called. Do not invent a wild-type person because a probe was absent.
- Genotype: the alleles present, at a site or a set of sites
- Homozygous: both copies match
- Heterozygous: the copies differ
- Not called: the assay did not measure the site
What phenotype actually names
Phenotype is the observable state: height, blood type, a lab value, a disease, a drug reaction, the presence or absence of a finding a clinician can name. It is the body in a world, not a printout.
Some phenotypes are almost categorical. ABO blood group is close. Many are quantitative: centimeters of height, millimoles of glucose. Some are clinical judgments with fuzzy borders. The word still holds. You are pointing at the trait, not the letters.
A file has no phenotype in it. A relative's story is a phenotype with recall bias. A clinic note is a phenotype with a date. None of those is the genotype, and none is the whole person.
The mapping is leaky on purpose
Does genotype equal phenotype? No. That is the whole point of having two words. Some mappings are tight. Some are so loose that the genotype is a weak hint.
Penetrance is the fraction of people with a genotype who show the trait. Incomplete penetrance is the usual scientific English for the leak. Variable expressivity is the leak in intensity: same genotype, different severity.
If a company treats a SNP as a yes/no identity, they have collapsed the mapping for a sale. Biology did not collapse.
Environment, chance, and the rest of the genome
A phenotype has causes besides the site you looked at. Other genes modify the effect. Age changes what is visible. Nutrition, infections, drugs, sunlight, and luck do work that no FASTA file records.
Identical twins start with nearly the same genotype and still diverge. That fact is not a slogan against genetics. It is evidence that genotype is not a screenplay.
Even a single-gene example leaks. Two people homozygous for HFE C282Y do not equally load iron. Two people with a BRCA variant do not equally develop cancer. The letters were real. The lives were not copies.
When the mapping is tight enough to be useful
Some genotype-phenotype pairs are tight enough that medicine uses them. Sickle cell disease follows homozygosity for a specific HBB allele in a way textbooks can draw. Some drug-gene pairs change dosing because the enzyme phenotype follows the genotype closely enough to matter.
Tight is not perfect. People with the sickle genotype still differ in severity. People with a pharmacogenetic genotype still need a clinician and a drug history. Useful is not the same as equal.
Blood type is the classroom favorite because the phenotype is a lab test with a short list of outcomes. Most traits you care about are not blood type.
Polygenic scores are not a phenotype
Most everyday traits are polygenic. Many sites, small effects, plus environment. A polygenic score sums some of those sites from a file and reports a number against a study population.
That number is not height. It is not a diagnosis of diabetes. It is not you. It is a statistical summary with ancestry limits, missing sites, and a ceiling on how much variance the score captures.
Treating a score as a phenotype is the same error as treating a SNP as a personality. Genotype in, number out, body still unmeasured.
A download is not a phenotype
People collapse the words because the file is what they have. A 23andMe export is genotypes at prechosen sites, on a stated assembly, with missingness. It contains no height, no diagnosis, and no family story.
A clinic note is the opposite object. It names findings in a body. It may mention a variant. The variant is evidence, not the phenotype itself. A relative's anecdote is a phenotype with recall bias and no assay.
Keep them in separate columns and most DNA-test confusion evaporates. Mix them and you will treat a SNP as a disease and a disease as a SNP. If a report prints a trait from a letter with no exam, no date, and no uncertainty, it printed a guess and labeled it a phenotype.
Reading letters without inventing a trait
If you have an rsID, you can look up the public record of that site: the gene, the alleles, a ClinVar label if anyone submitted one, a frequency in gnomAD if the site is there. That is still genotype-level annotation of a name, not your phenotype.
genome.sh does that lookup without uploading a genome. Matching the name to the two letters in a local file is a separate step. Inferring a trait from those letters is a third step, and it is the one most people skip badly.
A SNP is one kind of genotype difference, a single-letter site. Most of a life is not a SNP. If you want the molecular object, read the SNP guide next. If you want the clinical object, you need a clinician. This page is informational, not a diagnosis.
- Public lookup: what databases say about a variant name
- Local genotype: the alleles actually called in a file
- Phenotype: the trait in a person, measured in a life
- Diagnosis: a clinical conclusion genome.sh will not make
Questions
Does genotype equal phenotype?
No. Genotype is the DNA letters you carry at a place. Phenotype is the trait a body shows. Penetrance, other genes, environment, and chance sit between them.
What is genotype vs phenotype?
Genotype names alleles. Phenotype names the observable state. A recipe versus a meal is the analogy that stays accurate if you remember kitchens differ.
Can two people with the same genotype look different?
Yes. Incomplete penetrance, variable expressivity, age, environment, and the rest of the genome all leak. Identical twins are the existence proof.
What is penetrance?
The chance that a person with a given genotype shows the associated phenotype. Incomplete (or reduced) penetrance means some people with the genotype never show the trait.
What is an example of a genotype?
A/G at a SNP, or two copies of the HBB sickle allele, or heterozygous for HFE C282Y. It is a statement about DNA letters, with an assay behind it.
What is an example of a phenotype?
Blood type A, 178 cm of height, a fasting glucose, sickle cell disease as a clinical diagnosis, or no iron overload despite a risk genotype. It is a statement about a body.
Can environment change phenotype?
Yes. Nutrition, drugs, infections, sunlight, and a hundred unmeasured exposures change traits without rewriting the inherited genotype. Somatic DNA can also change in some tissues. That is still not the same as rewriting the germline.
What is the difference between a SNP, a genotype, and a phenotype?
A SNP is a site where a single letter commonly varies. A genotype at that SNP is the letters a person carries, if they were called. A phenotype is the trait, if anyone measured it. The three words are not interchangeable.