Biology 101 · Chapter 7

Genetics & inheritance

The trait was there the whole time — just hidden. You carry two copies of (almost) every gene, one from each parent. For eye colour, one common version builds brown pigment and a quieter version builds little. A parent can carry one of each and still look fully brown, because the brown version is enough on its own — the blue version rides along, silent. When two such parents each happen to pass the silent version, the child has two silent copies and nothing overrides them. The trait didn't appear; it was recessive, waiting for a copy with no brown partner to mask it.

Two copies, and which one shows

A gene is a stretch of DNA with a job (Chapter 2). An allele is a particular version of that gene — a spelling variant. Because your chromosomes come in pairs, you hold two alleles per gene. They can match or differ, and a simple rulebook decides what you actually see:

• Dominant allele (write it capital, B) — shows its effect even with only one copy.
• Recessive allele (lowercase, b) — shows only when both copies are recessive, with no dominant copy to mask it.

So BB and Bb both look brown; only bb looks blue. That gap — between the genotype (the alleles you carry) and the phenotype (the trait you show) — is the whole engine of the puzzle. A Bb person is a silent carrier: brown-eyed, but passing the blue allele to half their children on average.

One trap worth killing early: dominant does not mean common, strong, or better. It only means "shows up with one copy." Plenty of dominant alleles are rare, and plenty of recessive ones are everywhere. Dominance is about masking, not about winning.

Drawing the cross

To predict a child, you only need to know that each parent passes one of their two alleles, at random, to each child. A Punnett square just lays out every way the two parents' alleles can pair up. Cross two carriers — Bb × Bb — and the four equally likely combinations fall out:

Two things are worth pinning down. First, the 3:1 ratio is about probability per child, not a quota — four children could all be brown, or all blue; the dice have no memory. Second, this is exactly the hook resolved: two brown carriers, one-in-four odds of blue, the trait reappearing from copies that were there all along.

Why a trait can skip a generation

Because a recessive allele can be carried unseen, a trait can vanish for a generation and resurface in the next. Drawing the family as a pedigree makes the silent carriers visible:

Where the shuffle comes from

Why is every child a fresh draw, and why are siblings never identical? The answer is the special cell division that makes eggs and sperm — meiosis, the variation-making cousin of the copy-and-split from Chapter 5. Meiosis does two things ordinary division doesn't. It halves the chromosome count, so each gamete carries just one allele per gene (and the full pair is restored at fertilisation). And it shuffles: the maternal and paternal copies of each chromosome are dealt independently, and chromosomes even swap segments mid-division (crossing over). The result is that two parents can produce an astronomical number of genetically distinct children — the engine of variation that Chapter 6's natural selection then filters.

One honest caution before the toolkit. The clean "one gene, two alleles, 3:1" story is real but it's the simple case. Most traits you care about — height, skin colour, disease risk, and yes, real eye colour — are polygenic: shaped by dozens to thousands of genes plus environment, producing smooth ranges rather than tidy categories. The brown/blue model is a teaching scaffold and a good description of genuinely single-gene traits (like cystic fibrosis); just don't expect every trait to sort into neat squares.

Work one, then finish one

Worked: Two carriers, Bb × Bb. The square gives genotypes 1 BB : 2 Bb : 1 bb. Since BB and Bb both look brown and only bb looks blue, the phenotype ratio is 3 brown : 1 blue — so each child has a 1/4 chance of blue eyes and a 1/2 chance of being a brown-eyed carrier like the parents.

Your turn: Now cross a carrier with a blue-eyed partner, Bb × bb. What fraction of children are expected to have blue eyes? (Answer: the gametes are B,b from one parent and b,b from the other, giving Bb, bb, Bb, bb — so 2 of 4, a 1/2 chance of blue.)

Why this earns a place in your toolkit

Inheritance is where biology becomes a data discipline. Because traits trace to alleles, and alleles are just rows of letters (Chapter 2), predicting phenotype from genotype is a machine-learning problem. Genome-wide association studies scan millions of variants across millions of people to find which alleles nudge a trait; polygenic risk scores then sum thousands of those tiny effects to estimate someone's risk for a disease — a literal weighted model over your genome. The same genotype-to-phenotype map is what gene-editing tools like CRISPR aim to rewrite, and what drives breeding of crops and livestock. And the shuffle-and-select logic here is the direct ancestor of the genetic algorithms from Chapter 6: recombine, mutate, score, repeat. Read inheritance well and you've got the conceptual spine under modern genomics, personalised medicine, and a good slice of applied AI.