Switch Proteins: How a Cell Reads DNA ✎ express

Original: "In vitro binding energies capture Klf4 occupancy across the human genome"
arXiv:2601.16151 · 2026-01-22 · CC BY-NC-SA 4.0 · independent analysis · 1 min
Scientists measured how tightly the protein Klf4 grips different DNA sequences and learned to predict its behavior across the entire genome.

A cell is like a kitchen where thousands of recipes (genes) are collected in one thick book (DNA). The chef-protein Klf4 flips through the pages and decides which recipe to cook today. But how does it find the right line among billions of letters? Previously, researchers thought the protein looked for a short 'password'—a few consecutive letters. Now it turns out it senses much more.

Scientists took thousands of short DNA fragments and measured how tightly Klf4 clings to each. Even 'wrong' sequences are grabbed, just more weakly. It's as if the chef recognized not only the exact recipe but also similar pages with some letters smudged. The grip strength varies smoothly, not in a yes/no fashion.

The same protein can switch a gene on in one part of the cell and off in another. It all depends on how firmly it sits on the DNA.

Next, the researchers built a model that adds up the influence of each DNA letter like blocks in a tower. It turned out that letters don't act alone but together: neighboring letters can strengthen or weaken the protein's grip. They tested the model on long DNA strands stretched with optical tweezers and on the entire human genome. The predictions matched reality without any extra tweaking.

Now we can predict in advance where Klf4 will land in any cell, and what it will switch on there. This is a step toward understanding how cells remember their state (for example, why stem cells stay stem cells) and what goes wrong in cancer.

Klf4 is one of the four 'Yamanaka factors' that can revert ordinary cells back to stem cells. In other words, it literally knows how to rewind a cell's development.

In the series 'Westworld', android behavior is governed by 'cornerstones'—short program phrases. Proteins like Klf4 work similarly: short DNA stretches act as switches for entire cellular behavior programs.

Scientists
William Lawrence Bragg
Concepts
binding affinity genetic code predictive model gene expression amino acid sequence protein language model sequencing Fisher's fundamental theorem of natural selection Post-transcriptional regulation