The Heads of University Centres of Biomedical Science claims that the subpanel that will be assessing biomedical research is unfairly biased against the subject.
The group argues that, if the split of submissions is on a par with that of the research assessment exercise in 2008, an extra four biomedical scientists would need to sit on the panel for it to be representative.
Chas Chowdrey, president of the group and head of biomedical sciences at the University of Westminster, said the issue was not a "minor anomaly" that could be ignored.
"It is of such magnitude that failure to rectify it would undermine the credibility of the REF in a subject area of enormous importance for the health service and the medical biotechnology industry," he said.
"It is very disappointing that the REF team has not acted on our concerns...(it) is not only grossly unfair but will run the risk of serious challenges to the REF outcome."
A Higher Education Funding Council for England spokesman said the subpanels had "reviewed their membership at an early stage to ensure they had sufficient expertise to contribute to the main panel criteria. In response, the UK funding bodies made some additional appointments including an additional biomedical scientist," he said.
"It will be vital that panels have the full range of expertise to assess submissions and to carry the confidence of their communities in the REF outcomes.
"In the run-up...the REF team will survey institutions to gain an up-to-date picture of the likely volume of submissions in each discipline, and (then) appoint additional assessors to the panels as necessary."