AI Research Tool Bridges Two Labs in the Hunt for ALS Therapies

When MIT mechanical engineer Ritu Raman decided to study ALS, she faced a sprawling and contradictory research literature that would normally take months to master. An AI system called Co-Scientist compressed that timeline dramatically, helping her interrogate the evidence against her own experimental models — and ultimately pushing her into an unexpected collaboration with her husband, chemical biologist Ryan Flynn of Boston Children's Hospital.

The result is a joint research program, built around Co-Scientist's suggestions, that is now hunting for RNA-based mechanisms — and potentially RNA-based drugs — that could target amyotrophic lateral sclerosis, the neurodegenerative disease that progressively destroys voluntary muscle control.

Two toolkits, one disease

Raman and Flynn approach human biology from opposite directions. Raman, a mechanical engineer at MIT, builds living nerve and muscle tissues that model diseases affecting voluntary movement. Flynn, a chemical biologist at Boston Children's Hospital, maps RNA on the surface of cells, studying how that RNA influences cellular communication and how pathogens exploit it to invade.

Their toolkits rarely overlap. That is precisely why the ALS project required both.

Raman's move into ALS took her outside her usual domain. The field's literature is large, messy and often contradictory — the kind of body of evidence a new entrant can spend months simply cataloging before running a single experiment. According to the account of the collaboration, Co-Scientist changed that equation. The system helped Raman interrogate the existing evidence in relation to her tissue model, convert vague ideas into testable hypotheses, and rank potential research directions according to the trade-offs laboratories actually face: feasibility, cost, and potential risk–reward.

A catch that became a collaboration

Co-Scientist's strongest leads came with a complication. They pointed to processes at the surface of cells, where much of cellular communication is mediated. Raman could manipulate her engineered tissues and measure the outcomes, but decoding the molecular interactions driving those surface signals lay outside her expertise.

That gap became the catalyst for the partnership. Raman brought her new research directions to Flynn, whose lab specializes in exactly that territory — RNA on the cell surface and its role in how cells talk to each other.

The pair then used Co-Scientist iteratively. Rather than treating the system as a one-shot answer machine, they combined its best ideas into research pathways that united their distinct capabilities: Raman's engineered living tissues on one side, Flynn's RNA surface-mapping techniques on the other.

Why it matters

The collaboration illustrates a practical shift in how AI systems are being positioned in biomedical research. Co-Scientist did not replace either lab. It acted as a connector — compressing a literature review, surfacing hypotheses neither researcher would have prioritized alone, and flagging the exact point where one lab's competence ended and another's began.

For ALS research specifically, the stakes are concrete. The disease lacks effective therapies, and novel mechanistic targets are scarce. By steering the collaboration toward cell-surface RNA — Flynn's specialty — the system helped open a line of inquiry that neither lab would have pursued as rigorously on its own.

The teams' hunt is now on for novel RNA-based mechanisms that could be used to target ALS, with RNA-based drugs as a potential downstream outcome. Whether that search produces viable therapeutic candidates will depend on experiments still ahead — but the pathway itself was assembled, in part, by an AI system bridging two disciplines and two labs.