Every one of the ~20,000 genes in your body doesn’t just make one version of a protein. Cells are constantly making choices about which version of each RNA message to produce. One of the least understood of these choices is called alternative polyadenylation (APA): essentially, where a gene’s message gets “clipped” and finished off determines how stable it is, where in the cell it goes, and how much protein it makes. Errors in this process show up in cancer, neurodegeneration, and many other diseases.

The problem?

Aside from the core machinery, scientists have had very little idea which of the roughly 4,000 predicted RNA-binding proteins in the human genome were actually in charge of making those clipping decisions. So UCSD Professor Gene Yeo, Sanford Labs’ Chief Scientist and Co-Founder, and colleagues built a screen to find them. Their team systematically tested 879 human RNA-binding proteins to see which ones could flip this molecular switch. The result: 63 high-confidence regulators, the vast majority of which had never been connected to this process before.

Two of the surprise hits are particularly notable:

  • GRB2 — a protein long studied as a cancer signaling protein — turns out to have a completely separate, previously unknown life inside the cell nucleus, where it directly touches the RNA clipping machinery.
  • RNPS1 — known as a splicing factor — moonlights as an APA regulator by physically grabbing a different part of the same clipping complex. We also built an AI model fine-tuned on protein language that can predict which proteins are likely to be APA regulators just from their amino acid sequence alone, without any experiments. It correctly identified 13 out of 14 novel activators in a blind test. This is a powerful tool for virtual drug screening.

Why does this matter for medicine and biotech?

APA is increasingly recognized as a major driver of disease and a largely untapped lever for therapeutic intervention. Knowing which proteins control it and which protein domains are doing the work opens the door to designing molecules that modulate RNA processing in highly specific ways.

VIEW ALL LATEST NEWS