See more of the story

A U.S. military cemetery south of Rome has a grave that is thought to contain an Army private named Melton Futch. But the marble marker reads only, "Here rests in honored glory a comrade in arms known but to God."

It is one of about 6,000 graves of U.S. troops killed in World War II whom the military was not able to identify with the technology of the time.

Today, of course, increasingly sophisticated techniques make it possible to obtain a genomic profile that can reliably confirm identity. But in order to work, DNA identification requires a sample from a blood relative for comparison. And in the cases of many of the WWII dead, the military can find no siblings, no parents, no children, not even distant cousins.

So the Defense Department is considering trying a different approach: Instead of finding relatives and then matching their DNA, military researchers want to use the DNA to find the relatives.

It is a tactic that has helped solve cold murder cases. "The technology is there — we just have to develop the policy to use it," said Timothy McMahon, who oversees DNA identification of remains for the Armed Forces Medical Examiner System.

The cold-case DNA approach has the potential to solve cases that have stumped researchers for years, including that of Futch, 20, the poor son of a sawmill worker who had lied about his age to enlist at 16.

One cold night in December 1944, Futch wrapped himself in a green wool overcoat and crept toward a hill in Northern Italy, as part of a raiding party. The Germans were waiting.

The Americans fell back, but Futch was nowhere to be found.

After the war, locals stumbled on the bones of a soldier on the hillside. The pockets held Futch's address book and a letter from his wife. But what seemed like a straightforward identification soon unraveled.

Army grave-registration examiners could not match the teeth of the dead man to the private's dental records, and while the Army estimated that the bones belonged to a man who was several inches taller. The case was reopened a few years ago, but no relatives could be found.

Critics of the current approach — a plodding process that has yielded fewer than 200 identifications a year with a budget exceeding $150 million — say the government should obtain DNA samples from every unknown's remains and start running them through every possible DNA database.

"They are doing it backward, so you have policy getting in the way of science," said Ed Huffine, who headed testing of remains for the Armed Forces DNA Identification Lab.