Mapping the layer of biology where a lot of disease actually lives

Arc Institute, with Biohub and the University of California, San Francisco, posted an atlas this week of where proteins grab onto RNA.

It covers 67 cancer-associated RNA-binding proteins and maps more than 9 million binding sites at single-letter resolution. Humans have roughly 1,500 of these proteins, about 7.5% of all our protein-coding genes, and they decide which messages in a cell get read, held or destroyed.

The trick is consistency: every protein gets the same small tag and goes through one identical pulldown, with infrared detection instead of radioactive labels, so differences in the data reflect biology rather than method.

It's a preprint, not yet peer reviewed. It covers 67 of around 1,500 proteins, the tag could nudge how a protein behaves, and the work was done in human cell lines, so binding may differ in living tissue.

Think about the arc here. We sequenced the genome, the letters. Then we mapped where proteins bind DNA, the first layer of control. This is the next layer down: where proteins bind RNA, after the gene is read but before the protein is made. That middle layer is where a lot of disease actually lives, and it has been the least mapped of the three.

Each layer we make legible becomes the training ground for the next generation of models. This one opens a layer that was mostly dark.

Read the primary source →