A growing number of companies and academic centers have been developing the next generation of DNA sequencers - the machines that turn samples of DNA into readouts of A's, T's, C's, and G's.

Some of the efforts are starting to bear fruit, and a new study offers a glimpse of how powerful new DNA sequencers might benefit cancer patients and researchers.

The study found that a new sequencer developed by 454 Life Sciences in Branford, Conn., was better than conventional sequencers at detecting rare mutations associated with cancer in a sample of lung tumors.

"Tumors are really difficult to analyze because they have a lot of normal cells, and we are interested only in the cancer cells," says Dr. Michael Egholm of 454 Life Sciences.

Conventional sequencers, which employ a method pioneered by Dr. Frederick Sanger in the 1970s, take averages at each DNA location in a sample. As a result, rare mutations might be lost in the statistical noise and never show up.

454 sequencing appears to solve this problem. The method generates so many copies of each molecule that even a rare mutation can "stick out like a sore thumb" at the end of the day.

The company realized several years ago that the sequencer might benefit cancer patients by providing more accurate diagnoses of tumors. To test the machine, the company teamed up with Dr. Matthew Meyerson of the Dana-Farber Cancer Institute.

The study focused on the epidermal growth factor receptor (EGFR) gene, which is mutated in some lung cancers and is the target of several anticancer drugs.

Using the new sequencer, the researchers identified all of the known mutations in 22 tumor samples. But they also discovered several mutations overlooked by Sanger sequencers.

"We found the mutations we knew were there, and in the process discovered ones we did not know about," says Dr. Roman Thomas, a medical oncologist at Dana-Farber and the study's lead author.

Further experiments suggested that the newly discovered mutations were real and may have contributed to the cancers, the researchers report in the July Nature Medicine.

In one particular tumor, two EGFR mutations were found that were unknown when the patient was alive. The researchers suspect that one contributed to the disease while the other may have caused the patient's treatment to fail.

Dr. Thomas points out that reliable genetic information about tumors is critical if some patients are to receive the most appropriate treatment.

"You need a technology that allows you to provide a diagnosis for all patients regardless of the quality of tumor sample," he says. "The 454 method worked well in almost every imaginable type of specimen, including paraffin samples."

Dr. Egholm cautions that the study was small and that more research on the mutations is needed. "But the results are suggestive," he adds.

The technology could also be used to answer questions about cancer biology, including perhaps the role of EGFR mutations in lung cancer. Dr. Meyerson and others have reported that certain mutations can help predict the course of disease.

But not all studies have replicated the findings. One theory is that the negative studies failed to detect relevant mutations and more sensitive sequencing might lead to answers.

"We do not know what the answer is, but we believe that we have the technology to resolve the controversy by sequencing more deeply," says Dr. Egholm.

The 454 sequencer, which costs about $500,000, has been on the market since last year, and other high-throughput sequencing technologies are in development.

These could benefit efforts such as The Cancer Genome Atlas (TCGA) Pilot Project, which was launched recently by NCI and the National Human Genome Research Institute.

TCGA will use various technologies to identify aberrant genes for "re-sequencing." Beyond the pilot phase, the ultimate goal is to establish a catalogue of the major genomic changes in cancer.

By Edward R. Winstead