What if AI could help scientists notice a promising clue hidden in a vast collection of DNA sequences, then hand it over for careful laboratory work?
That is the interesting part of today’s biology story. There is also news of fresh funding for a London-based biotech company. Both are worth watching, but neither is a new medicine.
Hundreds of AI agents search DNA data
Anthropic says its life sciences researchers asked Claude to look for unusual reverse transcriptases, enzymes that copy RNA into DNA, in large DNA datasets.
With high-level direction from researchers, roughly 950 AI agents searched and analysed the data for about 21 hours. One spotted an unusual pattern of repeating DNA sequences beside a reverse-transcriptase gene.
Human scientists then investigated the lead in the laboratory. Their early experiments found that the repeat array produces distinct short RNAs. The researchers call the associated system ART, short for array-associated reverse transcriptases.
That sequence of events matters: AI helped identify a candidate in a huge dataset; people followed it up experimentally. Neither part alone would tell the whole story.
The researchers say the function of ART is not yet known. Some of its features resemble patterns seen in other biological systems, including CRISPR-related repeat arrays, but that does not make ART a new CRISPR tool, a proven gene-editing technique or a treatment.
The results have been shared as a preprint, and further experiments are underway. It is an intriguing starting point for research, not a demonstrated clinical advance.
Sources: Anthropic: research and laboratory follow-up ↗ · Research preprint ↗
London biotech attracts funding for future research
London-headquartered Basecamp Research has announced a $140m Series C funding round to develop its AI models for designing potential therapeutics and advance its research pipeline.
The company says it combines biological data with AI to investigate new approaches to medicines, including future cell therapies. It has described preclinical work and plans to move its pipeline towards clinical development.
That is an investment announcement and a commitment to future research. It is not evidence that the funding has already delivered an approved treatment or better patient outcomes.
For UK science, it is a sign of support for ambitious research. For patients, the important steps still lie ahead: careful testing, independent scrutiny and clinical trials where appropriate.
Source: Basecamp Research: funding announcement ↗
AI can find leads; people still have to test them
These developments are at different stages, but they point to the same practical question: where can technology remove a genuine bottleneck?
In one case, agents sifted through data that would be laborious to inspect manually. In the other, investment is backing further work on difficult biological problems. Neither removes the need for human judgement, laboratory experiments or evidence from patients.
Start with the problem. Use AI where it helps. Let scientists do the proving.
No AI for the sake of AI.
Less hype. More real-world AI progress.