In a small wooden hut at the Carnegie Institution's research station at Cold Spring Harbor on the north shore of Long Island, in the late summer of 1948, a forty-six-year-old maize geneticist named Barbara McClintock spent another long day looking down a microscope at the chromosomes of a corn kernel.
The orthodox view of her field, that summer, was that the genes on a chromosome had fixed addresses, the way books have fixed addresses on a library shelf. The address was the gene. The gene was the address.
McClintock's chromosome did not behave that way. The locus moved. It jumped from one position to another, in different cells, in successive generations. She named the moving element Dissociation. She named the regulator that moved it Activator. She had no theory for why a gene would move, because her field's theory did not allow it to.
In the summer of 1951 she presented her finding at the Cold Spring Harbor Symposium. The reception, as she later described it in her own correspondence, was "puzzlement, even hostility." For the next twenty years she stopped publishing in the major genetics journals. She continued the work. She told a friend that she would simply have to wait for the field to catch up.
It did, eventually. In the early 1970s the same kind of jumping element was found in bacteria. By the 1980s it was found everywhere: in fruit flies, in mice, in humans. About half of all the DNA in your body, it now turns out, is descended from the kind of mobile element McClintock was describing in 1948. She had not been describing an anomaly. She had been describing the largest single feature of the genome.
The Lasker Prize came in 1981. The Nobel came in 1983. She was eighty-one years old. She is the only woman who has received an unshared Nobel in Physiology or Medicine.
This is the story of how, last week, a different group of geneticists, at a different set of institutions, announced that they had found another seventeen hundred and eighty-five things in the same human genome that the field had been calling noise.
Some background on what was being called noise.
When the Human Genome Project finished, in 2003, the field had a parts-list of about twenty thousand protein-coding genes. The proteins were the working pieces of the cell. Everything else, which is to say about ninety-eight per cent of the DNA, was widely called junk. The phrase, coined in 1972, stuck.
It turned out the junk was doing something. In 2012 a large international consortium reported that as much as eighty per cent of human DNA had identifiable biochemical activity. Stretches of DNA that were not protein-coding were nonetheless being read, copied, switched on, switched off, used. Junk DNA was, like McClintock's anomalies, not what the textbooks had said.
There was a separate problem, hiding in the protein side of the same genome.
A ribosome, the cellular machine that builds proteins, does not always start reading at the place biologists annotated as the beginning of a gene. It can begin in unexpected places. It can read short stretches that the annotation software dismissed as too short to be a real protein. The proteins it produces in those short reads are typically only twenty to fifty amino acids long, well below the threshold the field used to define a protein at all. They were too small to be picked up reliably by classical mass spectrometry. They did not appear in the reference databases that geneticists used to interpret their experiments. They were, in the textbook sense, invisible.
On the seventh of May twenty twenty-six, in a paper in Nature titled "Expanding the human proteome with microproteins and peptideins," an international consortium called TransCODE published the first systematic catalogue of these short proteins. The team analysed roughly seven thousand two hundred of these dismissed reading frames across multiple human cell types, using improved mass spectrometry and a technique called ribosome profiling that lets you see, directly, which sequences a ribosome is reading.
They found firm evidence for one thousand seven hundred and eighty-five short proteins that had not previously been catalogued.
Because the field cannot yet say what each individual one of them does, the consortium proposed a new word. They called the new entries peptideins. A peptidein, by their definition, is a short amino-acid chain that does exist as a discrete molecule in the cell, but whose biological function is, for now, ambiguous. Some of the seventeen hundred and eighty-five they catalogued already have functions: a substantial subset are essential for cell survival, and several have been implicated in cancer. The rest are, for the moment, suspected of doing something or nothing.
The new entries will be added to the reference databases that biologists use. The known human proteome, by this single paper, has been expanded by approximately ten per cent.
The arc closes. The history of the genome is, repeatedly, the history of the field discovering that the noise was the signal. Mendel's laws were ignored for thirty-four years. McClintock's transposons for thirty. Junk DNA was the textbook view for forty. Now the dark proteome, the small proteins translated from reading frames the annotation software used to throw away, has joined the working list.
Each generation of geneticists has been confident about what part of the genome is doing nothing. Each generation has been wrong. The TransCODE paper is the latest correction. It will not be the last.
Barbara McClintock kept a small unfinished maize plot at Cold Spring Harbor until the year of her death in 1992, age ninety. She would, one suspects, have approved.