The research, published in the `Nature` journal, used a new technique which tested hundreds of genes at once, rather than one at a time.
The scientists at the Whitehead Institute for Biomedical Research said there was "a lot of potential for significant impact" if drugs could be developed.
In fact, they used small, disruptive, snippets of genetic material which can turn off genes. They injected cancerous cells with the snippets to investigate which genes were necessary for tumour formation and growth. They found that the gene -- PHGDH -- was highly active, far more than usual, in 70 per cent of tumours which did not respond to hormone therapies.
Over expression of the gene results in the chemistry a cancerous cell changing and is involved in the production of an amino acid -- serine.
The hope is that by identifying the gene which leads to some breast cancers, a drug can be developed which interferes with its activity.
"There is a lot of potential for a significant impact if a therapy targeting the serine pathway were found to be effective.
"However, as we do not treat any patients in our study, or develop any chemical inhibitors of the pathway, it would be very premature to predict the response in the general population," lead scientist Dr Richard Possemato said.