Pharma
Deployment record
In the tracker →- Organisation
- Incyte
- System
- Kosmos Target Assessment
- Stage
- Pilot
- Function
- drug target assessment
- Built on
- Edison Scientific
- 01Incyte is developing the Kosmos Target Assessment system in collaboration with Edison Scientific to enhance drug target selection processes.
- 02Kosmos can complete assessments in under one workday, significantly reducing the time required compared to traditional methods.
- 03The system integrates internal research records and biomarker data to improve the quality of target assessments by anticipating committee questions.
Show all 5Show less
- 04Edison has partnered with Pfizer to incorporate proprietary data into the Kosmos platform, enhancing its capabilities for pharmaceutical research.
- 05Decision quality and long-term pipeline productivity will be measured to evaluate the effectiveness of the Kosmos system beyond just speed.
Incyte
Follow Deployed on Google
Before a drugmaker spends years developing a medicine, it must decide which biological target is worth pursuing. Incyte, a biopharmaceutical company developing treatments for cancer and inflammatory diseases, is using an AI scientist to help prepare the evidence behind that decision.
Working with Edison Scientific, Incyte is developing Kosmos Target Assessment to bring scientific evidence and strategic considerations into the reports reviewed by its target-selection committees. Edison recently reported that the system completed assessments in less than one workday on average.
“Kosmos produced in a morning what would have taken me 3 months of work to complete,” said Pichai Raman, senior director, computational biology at Incyte.
Looking Beyond the Biology
A target can have a strong biological rationale and still raise questions about safety or how it fits the company’s pipeline. Kosmos is intended to bring those considerations into the assessment earlier, helping scientists anticipate questions from the review committee.
Incyte announced its collaboration with Edison in May, starting with target discovery and validation, alongside translational biology.
“Our vision is for our data to become a learning system that enhances every decision,” said Pablo J. Cagnoni, president and global head of research and development at Incyte.
Edison spun out of the nonprofit research lab FutureHouse in 2025. Its Kosmos platform connects to internal research records and biomarker data, allowing researchers to draw on findings from previous work when assessing a new question.
Learning From Previous Reviews
Edison evaluated Kosmos against human-prepared assessments of the same targets. Both were scored using criteria drawn from transcripts of Incyte’s review meetings.
The week's drug discovery, trials and health-system deployments. Wednesdays.
The largest reported improvements were in strategic considerations and portfolio fit. Improvements in biological coverage were smaller, reflecting the expertise already present in the scientists’ assessments.
In a separate exercise covering six targets with known committee decisions, Kosmos proposed follow-up experiments that exactly or conceptually matched committee requests in 83% of comparisons.
The findings suggest the system can anticipate questions raised in previous reviews. The evaluation was retrospective, so it does not show whether Kosmos improves decisions on new targets or changes which programs receive funding.
Pfizer Brings Another Set of Research Data
Edison’s agreement with Pfizer, announced last month, extends its work into another pharmaceutical research organization.
Kosmos will combine public evidence with Pfizer’s proprietary data. Edison scientists and engineers will work alongside Pfizer teams to build the system into specific research workflows.
“Pfizer has spent decades generating data about what works and what doesn’t work in human biology,” said Sam Rodriques, CEO of Edison Scientific.
Working with that data also requires controls over where it goes. Edison’s technical account describes isolated computing environments and deployment within partner infrastructure, designed to limit data leaving the organization.
Scientists remain responsible for checking the evidence and deciding which experiments to pursue. Incyte and Edison have said they will measure decision quality and long-term pipeline productivity, extending the assessment beyond how quickly Kosmos prepares a report.
