Introduction
Count Mendel's peas with him. For stem length he grew 1,064 plants in the second generation: 787 long, 277 short. That is 2.84 to 1. Close enough to three to one that school biology has taught the clean version ever since.
Now try the same trick on height. How many genes would you guess? A Nature paper in 2022 pooled 5.4 million genomes and found 12,111 variants, scattered over about a fifth of the genome. None of them is "the height gene". Each nudges the total by a hair. And those variants explain 40% of the variation in European-ancestry groups but only 10 to 20% elsewhere, so even the big number is provisional. Most disease risk looks like height, not like peas.
History and Development
Gregor Mendel grew roughly 28,000 pea plants and published the result in 1866. His insight was arithmetic: traits came in discrete units that sorted independently, in ratios you could count. Nobody noticed. The paper sat unread for 34 years until 1900, when Hugo de Vries, Carl Correns and Erich von Tschermak each arrived at the same conclusions separately and found Mendel had got there first.
1953 gave the units a physical form, when Nature published the DNA double helix. Watson and Crick assembled that model from X-ray diffraction data produced at King's College London by Rosalind Franklin and Maurice Wilkins, whose papers appeared alongside theirs. Half a century later the Human Genome Project was declared complete in 2003. What it delivered was a sequence, not an explanation; working out what the sequence does is the project still running.
Core Concepts of Genetics
Four letters make up the whole alphabet. The double helix matters for mechanics, not elegance. Two complementary strands mean the molecule carries its own backup. Replication unzips it and reads each half.
Sequence becomes protein in two steps. Transcription copies a gene into RNA. Translation reads that RNA three letters at a time and strings amino acids together. Now the part people miss. Every cell in your body holds an identical genome. A neuron and a liver cell differ in one respect: which genes they are reading.
Inheritance is about dosage. Autosomal dominant needs one faulty copy, so the condition appears in every generation. Autosomal recessive needs two. It skips generations and surfaces in families with no history of it. A pedigree chart tells a counsellor what a symptom list cannot.
Subtopics in Genetics
Population genetics counts alleles rather than individuals. A frequency shifting across generations is evolution, measured. Genetic drift moves those frequencies by chance alone, and in a small population chance beats selection.
Polygenic traits are the ones Mendel could not handle. Height, yield, blood pressure. Hundreds of genes, each contributing a little, plus the environment.
Genomics reads whole genomes instead of single genes, and gene mapping locates a variant on a chromosome. Mutations supply the raw material for all of it.
Gene therapy is the applied end. One faulty gene, one disease, is the rare clean case. Most risk is a smear across many variants.
Applications of Genetic Research
Diagnosis came first and is still where genetics does the most good. Knowing which variant causes a child's condition ends the diagnostic odyssey families otherwise spend years on, even when no treatment follows. Gene therapy goes further and works best on the narrow cases: one gene, one tissue, a target you can reach. Sickle cell and some inherited blindness qualify. Most conditions do not.
Agriculture has been running this experiment longest, since every domesticated crop is a genetically modified organism in the literal sense, selected over millennia. What changed is precision and speed. Conservation uses the same tools backwards, measuring the genetic diversity left in a small population, because a species can be numerous enough to look safe while being too inbred to adapt.
Current Trends in Genetics
Personalized medicine has one real success and it is narrower than the brochures suggest. Pharmacogenomics. Some people carry a slow version of a liver enzyme, so a standard dose builds up in them and poisons them while clearing normally in everyone else. Test first, adjust the dose, done. That works now.
The grander version, your whole genome steering your care, keeps running into a dull problem with the reference data. Most of it came from people of European ancestry. Build a risk score on that and it predicts poorly for everyone else, which makes the technology least useful where health systems are already worst.
Then there is synthetic biology, which wants to treat genes as catalogue parts. Order the components, assemble, get the behaviour you specified. Cells do not cooperate. Move a working circuit into a different cell line and it misfires; run it for forty generations and the population quietly mutates your design away. We can read a genome for the price of a laptop. Writing one that does what the drawing says is another matter.
Further Reading
For deeper study, explore these resources: