The context rules

This September, Google DeepMind presented the AlphaGenome Atlas, a database with predictions for each of the nearly nine billion possible single-letter changes in the human genome. 98% of the genome does not encode proteins, but it largely regulates when and where genes are activated, and it is precisely here that it is hardest to read.

One of the first applications has been in rare diseases. In the case of a person with severe epilepsy, out of the four or five million variants in their genome, the Atlas placed a variant of the DNM1 gene at the top of the list. This variant creates a splicing site (splicing) that only acts in the brain and lengthens the protein in question by thirteen amino acids. RNA studies conducted from blood samples had clarified nothing, and now it is understood why: in blood, that part of the gene is barely expressed. A laboratory experiment confirmed the mechanism, and the researchers recommend reclassifying the variant as “likely pathogenic”. The authors themselves also warn: AlphaGenome is a research tool, not a clinical diagnostic tool; it is still missing many cell types; and the work has not undergone any external review.

If context decides what a letter means, it seems it can also decide the age of an organ. A Harvard team, led by Jesse Poganik, has measured the biological age of transplanted hearts, in mice and in 11 patients, with epigenetic ‘clocks’. These clocks read the chemical marks (methylation) accumulated on DNA. The results suggest that young hearts age rapidly within older bodies, and that old hearts rejuvenate within young bodies, at least according to these markers. If confirmed in more patients, it could be possible to accept older donors and ease waiting lists because of which many people still die. Even so, this study has not yet undergone external review, and the biological mechanism behind it is far from being understood.

However, the context is not only biological. We already spoke about this in this space: the next step for CAR-T therapies is to reprogram immune cells within the patient themselves (in vivo), without having to manufacture them outside. According to Nature, China is leading this race, above all, thanks to its system of investigator-initiated trials, which are more agile and cheaper than conventional ones. This September, the American FDA authorized for the first time the start of a clinical trial with an in vivo CAR-T developed in China. The same technology advances at a different pace depending on where it is tested. Here, a balance must be found: what speed is prudent when it comes to reprogramming the immune system? And who will have access to it first?

In the three cases we are discussing today, the place where something happens is as important as the thing itself. In science, context is also part of the information. If we do not know where a result was obtained, in how many people and with what rules, we only have a part of it; the other part is knowing how far it reaches, why it is important, and to whom it is addressed.