A Medical School research team last week received federal approval for a controversial project that involves gene splicing experiments with deadly diphtheria toxin.

The experiments potential benefits to cancer research outweigh the potential dangers of the project, which uses recombinant DNA techniques to introduce diphtheria-producing genes into E colt bacteria.

Dr. John R. Murphy, associate professor of Microbiology and Molecular Genetics, who heads the research group, has speculated that successful manipulation of the toxin gene could result in the development of a treatment for malignant melanoma a type of skin cancer.

The E coil bacteria are normally present in the human digestive tract, and some scientists fear experimental cells escaping from the laboratory could infect otherwise healthy people.

Containment

The dangers associated with the project are expected to be minimized by conducting the experiments in secure laboratories, originally designed for use in germ-warfare testing at Fort Detrick in Frederick. Md Additionally, the strain of E. coli being used in the research is a special breed which does not survive well outside a laboratory.

Although the experiments may prove un workable because of differences in the biology of the bacteria and the diphtheria toxin. Murphy has expressed hope that the project will demonstrate the feasibility of using a foreign host to produce toxin.

Murphy's previous work with the diphtheria toxin, one of the most dangerous persons known, has been considered safe because the fragments of the toxin molecule with which he works are biologically inactive.

The project's conditional approval from the National Institutes of Health will be published shortly in the agency's Federal register, a government spokesman said last week. Final government approbation should come after a 30 day public comment period.

Because work can not begin until after the comment period, an official starting date for the project has not been set. Murphy said last week that conclusive results could not be expected for at least a year.