Beneath the waters of Saipan's harbor, an upside-down warplane, its metal exterior now adorned with coral growth, sits quietly as vibrant tropical fish dart through its fuselage and engine. Slowly blending into the surrounding reef, this aircraft still tells a story decades after it fell into the sea.

The aircraft, known as a Grumman TBF Avenger, carried three crew members when it likely crashed during or after the 1944 Battle of Saipan — a crucial World War II campaign as U.S. forces advanced across the Pacific. Saipan, the largest of the Northern Mariana Islands, was a strategic goal, and its capture marked a major turning point. Of the crew aboard, only one survived. The remains of the other two have never been recovered but are not forgotten.

Today, this submerged Avenger has become a focus point for an advanced scientific effort to determine if DNA, even decades old, can be found at underwater crash sites. Scientists, working alongside the Defense POW/MIA Accounting Agency (DPAA), are exploring new technology that detects environmental DNA (eDNA) in surrounding water and sediment. The goal is to create a new method to trace the presence of missing service members — and potentially locate many more who vanished at sea.

The DPAA, responsible for recovering U.S. troops lost abroad, is especially interested in whether this eDNA technology can accelerate searches. Over 40,000 American service members are still missing at sea since World War II.

“We face tough challenges when investigating underwater,” explained Jesse Stephen, leading innovation at the DPAA. “Unlike land-based search efforts, underwater searches often face poor historical records, vast and hard-to-navigate environments, and difficulty recovering remains separated from wreckage.”

Stephen added that they wanted to test whether eDNA could be an effective method to flag the potential presence of human remains before full excavations begin.

As part of a major pilot program, the Avenger wreck is one of several sites selected for study. Other locations include Lake Huron, in the Great Lakes, and waters near Palermo, Italy. Overall, a dozen ship and plane wrecks—some military and others civilian—were analyzed. Some sites are of direct interest to DPAA, while others serve as controls or test comparisons in varying underwater conditions.

The initial results of the study show potential, although they also raised unexpected questions researchers didn’t anticipate.

eDNA as a Tool for Detecting the Unseen

The method involves detecting invisible traces of DNA left in the environment by living and deceased organisms. It's a relatively recent breakthrough that has become useful in fields such as conservation science.

“Without touching or collecting remains, we can potentially detect whether a person or animal had been present at a site, just from the DNA they shed into water or sediment,” said marine biologist Kirstin Meyer-Kaiser of the Woods Hole Oceanographic Institution in Massachusetts. She partnered with DPAA to explore this method in challenging marine environments.

eDNA first gained recognition when researchers used it in 2008 to find invasive bullfrogs in France, after traditional observation failed. Now, the technique is routinely used to identify numerous aquatic species. Meyer-Kaiser had the idea of using it as a kind of forensic “bone sniffer” to uncover DNA left by decomposed human remains — a possible improvement over slow underwater excavations that the DPAA typically conducts.

On land, eDNA studies have successfully recovered ancient DNA from sediment, often in environments such as dark caves where preservation is ideal. The ocean poses new challenges, with currents, fluctuating temperatures and varied depths complicating DNA detection and preservation.

A Detailed Look at the Avenger Aircraft Wreck

To test the approach, samples were taken in 2022 and 2023 from both sediment and water at seven submerged planes — mostly from WWII — and five shipwrecks across multiple environments: off Palermo in the Mediterranean, a section of Lake Huron, and two locations around Saipan including the lagoon and deeper sea.

By comparing sites in fresh versus salt water, warm versus cold, and shallow versus deep, the team hoped to understand where eDNA would be most effective for identifying human remains.

Only some of the selected locations were tied to missing military personnel. The rest served as controls to gather clean comparisons. At nearly 300 meters deep, some Saipan sites could only be reached using remote-operated underwater vehicles. Others, including the Avenger wreck, lie within scuba diving range. Calvin Mires, an archaeologist from WHOI, oversaw sampling operations at these locations.

Pushing the Boundaries of Science

An experienced underwater archaeologist, Mires collected samples at all three test areas. “This method is outside my usual work,” he admitted. “But this kind of unknown is exactly why I chose science.”

At the Saipan lagoon site, Mires and his team used waterproof tablets displaying detailed maps to identify sampling locations. To prevent sample contamination, all team members were DNA-profiled to filter out their own genetic material. During collection, they wore gloves and hoods when possible.

Sampling water was straightforward, with divers using Niskin bottles — specialized gear designed to trap water at set depths. Sediment sampling, however, proved tricky. Standard collection tubes struggled in the coral-covered wreckage, forcing Mires to carefully use trowels where tubes wouldn’t work.

Despite challenges, samples were collected—though their journey to lab analysis was almost derailed. Initially meant to be transported on dry ice at -80°C, the team had to improvise when the airline refused to ship with dry ice. Instead, the samples traveled to Guam packed in regular ice and stayed in a freezer for about a month before reaching their final destination at the University of Wisconsin Biotechnology Center.

Fortunately, according to lead researcher Charles Konsitzke, the temperature hiccup didn’t compromise the results or analysis.

Demonstrating the Concept

Throughout much of 2023 and into 2024, researchers analyzed the material using a method called metagenomics – which sequences all DNA present rather than searching for specific species.

This approach helped identify not only human DNA, but also microorganisms that might play a role in decomposition, serving as indirect indicators of human remains nearby.

Meyer-Kaiser said, “We were looking for both direct evidence of human DNA and microbial signs that hint at human decomposition.”

One of the study’s key challenges was recognizing aged DNA amidst newer fragments possibly introduced by people recently visiting wreck sites. Newer DNA strands are usually longer, while older DNA tends to break into much smaller pieces — a way to identify age based on sequence length.

Researchers noted the appearance of many short human DNA fragments — some just 40 base pairs long — which supported the idea that older DNA could survive underwater. Longer DNA strands, by contrast, suggested more recent origins.

Though identifying ancient DNA helps, the rate at which it breaks down varies based on environmental factors, making it difficult to link DNA to a precise time period.

Elena Irene Zavala, a forensic geneticist uninvolved in the study, noted that DNA often binds to minerals in sediment, aiding preservation. However, degradation happens in unpredictable patterns — another variable the science is still exploring.

Interestingly, sediment samples proved far more informative than water samples. Meyer-Kaiser believes this relates to stability: “Water is constantly in motion and mixes easily, but sediment layers stay much more undisturbed, preserving information better.”

In sediment samples from Saipan and Palermo wrecks, researchers saw a significant concentration of older human DNA fragments, especially in areas suspected to contain human remains. In some samples, just one sediment core stood out — suggesting that the method could potentially zero in on DNA “hotspots.”

“We’ve confirmed the overall idea. We can detect and tell apart older human DNA from new,” said Meyer-Kaiser. “While we couldn’t verify these with actual remains, our approach shows we’re headed in the right direction.”

Questions Remain

Since no human bones were recovered during this study, scientists could not confirm a direct match between old DNA and missing service members. Meyer-Kaiser’s final report calls for full excavations at three promising sites to confirm whether DNA findings actually indicate human remains. It remains uncertain whether such digs will be approved.

“To verify these results, excavation is crucial,” she said. “Without it, we don’t yet have definitive proof.”

Konsitzke echoed her thoughts: “We need side-by-side recovery data to validate the method.”

Unexpected results at deepwater sites in cold conditions raised further questions. In both Lake Huron and the deep Saipan test beds, control sites — where no remains were expected — showed higher amounts of old human DNA than known crash sites. This contrasted with the warm-water locations, where patterns aligned more logically.

“It completely confused me that we found more DNA where we didn’t expect it,” Meyer-Kaiser admitted. Possible explanations could include runoff, decades-old sewage, or naturally shed human DNA accumulating even in remote locations.

Konsitzke added that colder waters may be enabling longer DNA preservation, and that even distant human activity could cause trace contamination.

Among other findings, researchers also detected microbial species. Some genes tied to skin and connective tissue were discovered near Palermo’s aircraft crash site, which Meyer-Kaiser hopes to explore further.

Looking Ahead

Jennifer McKinnon, a maritime archaeologist who has extensively studied WWII wrecks near Saipan, praised the effort, despite not being involved in the research. “It’s promising to apply new technologies to unsolved cases,” she said. “If this method helps missions like the DPAA’s, it’s worth pursuing.”

What happens next is still uncertain. Meyer-Kaiser hopes to start a second phase of the study at crash sites closely tied to known missing individuals. She’s also interested in researching microbes from the study to see if they could consistently signal the presence of human remains.

Stephen from the DPAA said the results will be reviewed carefully, alongside other pilot technologies like underwater drone imaging partnered with artificial intelligence.

He sees significant promise in the eDNA research but tempered expectations. “This isn’t about identifying individuals,” he explained. “It’s an investigative tool — a way to confirm whether an area merits further excavation.”

“We’re encouraged,” Stephen concluded. “Recovering degraded DNA is a major step. Each discovery brings us closer to knowing what’s possible.”