After six years of debate and thousands of gene-splicing experiments, a showdown vote on whether to abandon federal regulation of genetic engineering experiments is expected at the National Institutes of Health today.

Last September, the NIH's Recombinant DNA Advisory Committee gave its preliminary approval to abolition of the regulations. But since that vote, public pressure has put the outcome in doubt again.

Today's meeting, set for 9 a.m. at the Bethesda Marriott Hotel, is expected to open with several hours of debate. At the last meeting, a procedural vote that was taken as the best test of how the committee would vote came out 12 to 9 for ending regulation.

But two new members have since joined the committee, and others are believed to be reconsidering their positions. A change of two votes would alter the outcome, which will directly affect genetic engineering experiments at virtually all U.S. universities, and will indirectly affect the way such experiments are carried out at the booming commercial gene companies as well.

The technology in question, created about nine years ago, allows the mixing of genes from two or more species. Biologists originally feared that putting genes from one kind of microbe into another might also transform a harmless bug into a disease-causing one.

"There is no justification left, if there ever was one, no scientific justification for recombinant DNA rules," said David Baltimore of Massachusetts Institute of Technology, who is proposing an end to regulation. He said that after thousands of actual recombinations of genes without accidents, and some experiments designed specifically to test whether gene splicing is dangerous, most scientists have concluded that their worries about the new technology were groundless.

But another biologist and member of the committee, Richard Goldstein of Harvard University, said he was "not convinced there aren't any more risks." As an example, he named experiments that might take the genes for making biological poisons, such as botulism toxin, and put them into an organism that never before could make such poison.

But overshadowing scientific considerations in the last two committee meetings has been debate about the social and political consequences of ending the government's oversight of gene engineering.

Two chief proposals will come before the committee today. One, approved at the last meeting, would end mandatory federal regulation of gene splicing, and put the work back on the same footing as other possibly hazardous experiments in biology--with only a voluntary code for handling dangerous organisms.

The second proposal, which was previously rejected, would relax the guidelines by making all but a few experiments free from prior review by NIH. But penalties, which can end federal funding for a biologist or a university found violating the rules, would be retained.

The federal regulations were first adopted in 1976. They prohibited some kinds of experiments altogether, such as those with microbes that cause deadly diseases. They required that other, supposedly dangerous, experiments be done in specially built laboratories or with equipment to guarantee that the micro-organisms are confined.

In every committee meeting of the past few years, the federal rules have been relaxed more and more, to the point now that more than 95 percent of gene-engineering experiments no longer require any special equipment or NIH approval.