Copyright Cable News Network
A warplane encrusted with coral lies upside down on the seafloor of Saipan’s harbor, its landing struts thrust awkwardly upward as colorful fish weave in and out of the exposed fuselage, engine and gun turret. The wreckage is slowly being subsumed by the reef in the shallow blue waters of the tropical lagoon, but its story is far from over. Three crew members had been aboard the Grumman TBF Avenger as it went down, likely during or after the 1944 Battle of Saipan, when US forces launched an assault to secure the Northern Mariana Islands during World War II. American military officials viewed Saipan, the largest island in the Mariana archipelago, as a key base for operations in the Pacific, and the battle marked a pivotal victory for the US. Only one of the downed plane’s crew, however, walked away from the conflict; the remains of the other two men are presumed submerged, neither recovered nor forgotten. Now, the Avenger wreck is one focal point for scientists in a cutting-edge endeavor to understand whether traces of decades-old DNA can be preserved and detected at wreck sites. Working with the US Defense POW/MIA Accounting Agency, the researchers are testing innovative technology that uses environmental DNA, or eDNA, found in soil, sediment and water as a biological scouting mechanism. If successful, the method could help locate the remains of the missing service members — and perhaps thousands more. The Defense POW/MIA Accounting Agency, or DPAA, is tasked with recovering the remains of the more than 40,000 American troops presumed lost at sea since World War II. Officials are hoping eDNA will speed up the search. “We’re looking for ways to meet some of the challenges that are inherent in underwater investigations … some of the most challenging sites for us,” said Jesse Stephen, an archaeologist who is the DPAA’s chief of innovation. Investigating an underwater site, he said, tends to be more complicated than looking for missing service members on land for several reasons. There is typically a dearth of historical information about the conditions in which the vehicle went missing, and locating a sunken aircraft or a large ship in the vastness of the ocean is difficult and costly. Even when a site has been located, it’s hard to find human remains, which often become separated from a wreckage due to impact force or the movement of water. “We wanted to explore how you could you use eDNA as a general site investigation tool specifically to determine the likely presence or absence of human remains,” Stephen said. The decaying Avenger is one of a dozen wrecks in three very different underwater environments — the western Pacific Ocean off Saipan, Lake Huron, shared by Michigan and Ontario, and the Italian Mediterranean — investigated as part of the ambitious pilot project involving eDNA. Most, but not all, of the wrecks involved were sites of active interest for the DPAA. Initial results shared with CNN offer a glimmer of hope that the technique could help identify lost troops’ remains, but the findings have also ignited some new and puzzling questions. Environmental DNA as ‘bone sniffer’ The underwater clues that researchers are grasping for in this effort are invisible, almost ephemeral, but hold promise just the same. Environmental DNA detection is an exciting and relatively new approach that allows scientists to recognize and amplify the genetic material that all creatures, living and dead, shed into their environments, and it already has gotten results in some fields of research. “Without physically collecting an animal or a plant or human remains, you can determine whether something is there — ideally, where it is located — based on just the DNA that’s getting shed into the water or the surrounding sediment,” explained Kirstin Meyer-Kaiser, a marine biologist and associate scientist studying seafloor ecology at the Woods Hole Oceanographic Institution (WHOI) in Massachusetts, who partnered with the DPAA on the project. Until now, researchers have primarily used the technique in conservation efforts. It was first successfully employed in an aquatic environment in 2008 to detect the presence of an invasive species — the American bullfrog — in France, after more traditional methods of DNA detection, such as audio or visual surveys, had been unsuccessful. The approach has since been applied to identify a wide range of species, small and large, particularly in water environments. Meyer-Kaiser envisioned harnessing eDNA as a “bone sniffer” to detect and locate fragments of DNA left by decomposing human remains, which, if it were to prove feasible, would be a transformative alternative to the DPAA’s traditional approach: painstaking, monthslong underwater excavations to try and identify the bones or belongings of a lost pilot or other crew member. Environmental DNA has been successfully used by archaeologists to identify and extract human DNA, sometimes tens of thousands of years old, from sediment on land. However, the successful samples have largely been collected from dark caves or other cold and stable environments where DNA is preserved particularly well. The ocean, with its variable currents, depths and temperatures, is a starkly different environment. The ambitious project planned by Meyer-Kaiser involved the collection of water and sediment samples from seven downed planes, the majority of which were from World War II, and five shipwrecks in three locales: off the coast of Palermo, Italy; Lake Huron’s shipwreck alley; and two locations in Saipan’s waters, one the shallow lagoon where the Avenger crashed and the other deep ocean. The sampling took place in 2022 and 2023. The researchers chose wrecks located in a range of environments to help them understand whether eDNA might be a more effective mechanism to identify human remains in some conditions than others: salt versus fresh water; cold versus warm temperatures; shallow versus deeper depths. The military sites included were in Saipan and Italy, while the Lake Huron sites were commercial shipwrecks in relatively accessible freshwater locations. The experiment also analyzed wrecks at the same sites that are not associated with those who have gone missing to ensure that control samples were collected. The four wrecks at the deepest site off the coast of Saipan, some 300 meters (980 feet) beneath the surface, were only reachable by an underwater remotely operated vehicle, but others exist at depths that are easily accessible to scuba divers. Calvin Mires, a maritime archaeologist at WHOI and part of the DPAA project, was tasked with recovering the water and sediment samples using the methodology set out by Meyer-Kaiser. To the cutting edge A veteran of large-scale underwater excavations, Mires suited up to collect samples at the shallower Saipan site, where the Avenger is located, and supervised the process at Lake Huron and in Italy. “EDNA is a little bit out of my wheelhouse,” Mires said. “But this is what I got into science for, it’s going to that cutting edge, not knowing the answer.” In Saipan’s turquoise waters, Mires and his colleague referred to an annotated map on a waterproof tablet as they marked out collection locations over the wreck site using red, numbered flags. To avoid contamination, team members had their DNA genotyped so their own sequences could be identified and removed from the samples. The researchers also wore hoods and gloves when possible during field ops. Water sampling went smoothly, with the help of Niskin bottles, oceanographic tools for sample-collecting that consist of thick plastic tubes with stoppers on both ends. Researchers hold the tubes open during a dive before releasing to close them around a parcel of water. Sampling of seafloor sediment proved to be more difficult. Handheld push cores, clear-plastic tubes with one-way valves that collect sediment, caused unanticipated difficulties — particularly at the Avenger site, where the bony coral reef made it tough to collect fine sediment. In some cases, Mires used a small trowel to collect samples. His collection efforts were successful, but the samples nearly didn’t make the long journey back to the University of Wisconsin Biotechnology Center, where they were to be analyzed. The team had planned to store and ship the samples in dry ice to keep them at the required temperature of minus 80 degrees Celsius (minus 112 degrees Fahrenheit). However, the shipping company told the researchers at the last minute that the airline would not ship dry ice. The team found a solution: transporting the samples to Guam in regular ice, where they were stored in a freezer for a month before making the trip to the US. After an agonizing wait, Charles Konsitzke, the Biotechnology Center’s associate director and team lead for University of Wisconsin–Madison’s Missing in Action program, was able to confirm that the dry ice drama fortunately had no net effect on the quality of the data or interpretation of the findings. ‘Proof of concept’ Analysis of everything the divers gathered took place over the course of several months in 2023 and 2024. The team deployed a relatively new technique known as metagenomics, which involves sequencing all the DNA in a sample at the same time rather than simply searching for the species of interest. In addition to directly recovering human genetic material, this process allowed Meyer-Kaiser and her colleagues to potentially identify a proxy for human DNA: microbes that may be involved in the decomposition of human remains. “We sequenced all DNA that was in our samples, human or non-human. With the microbial data, we were essentially fishing for anything that might be associated with human remains,” she said. Pivotal to the study was the ability to distinguish shorter pieces of decades-old DNA, which might belong to a lost service member from longer pieces potentially left by a swimmer at a site a day before. Living human DNA consists of approximately 3 billion base pairs, arranged into long strands — chromosomes that are made up of genes. After death, human DNA breaks down into ever smaller fragments. As a rule of thumb, the shorter the sequence or read of base pairs, the older the DNA, Meyer-Kaiser said. The team measured the length of each human DNA fragment in the samples by calculating the number of base pairs along each strand. Generally, “old” or “degraded” DNA is considered to be anything less than 150 base pairs, Meyer-Kaiser said. The process took longer than expected because of some confusing signals in the data, Meyer-Kaiser added. The research was, however, successful in accomplishing one of its key goals: detecting and differentiating between contemporary human DNA and older sequences in the samples. In the samples, the scientists found peaks of sequences that were 40 base pairs in length, suggesting the human DNA had been there for a longer period of time. The other human DNA in the samples was more than 150 base pairs in length, suggesting a more recent origin. However, it’s not currently possible to pinpoint with any accuracy how long the DNA had been there because DNA fragments at different rates depending on conditions of the surrounding environment. Finding short fragments of DNA less than 40 base pairs long is an initial positive indicator of older DNA, but other environmental factors might be at play, according to Elena Irene Zavala, an assistant professor in forensic genetics at the University of Copenhagen. Zavala has applied techniques developed in the field of ancient DNA research to help identify unknown human skeletal remains from World War II and the Korean War. Zavala wasn’t involved with Meyer-Kaiser’s pilot study and hasn’t investigated aquatic cold cases. Her current research focuses on trying to understand how human environmental DNA preserves and degrades in sediment through time — on both long and short timescales. “DNA binds with minerals,” she said, and that “is how it can be preserved for thousands and thousands and thousands of years. If that degradation happens quickly or continues to be progressive through time, is still something that we’re trying to understand, and we see evidence of both.” The sediment samples revealed more information than the water samples, which did not surprise Meyer-Kaiser and her team. “Water moves around and, you know, water is impacted by wave motion, and there’s rain water, and it’s going to be a lot more ephemeral,” she said. “Whereas something that’s buried under the sediment, the sediment isn’t getting turned over nearly as quickly as water would be.” In particular, sediment samples from two aircraft in the Saipan lagoon and the Italian crash site contained a “high abundance” of shorter human DNA fragments in areas suspected to contain human remains. And at each of these sites, one sediment core had a much higher abundance of human DNA than other samples. This indicated that these sediment cores could indeed identify areas where human remains were located at a site, Meyer-Kaiser said. “We do have proof of concept. We are able to find human DNA. We are able to separate older from more contemporary human DNA,” she said. “We don’t have a positive control, but we know that collecting the DNA in this way, sequencing it in this way, analyzing it in this way, gives us the type of information and data that we’re looking for,” Meyer-Kaiser added. New puzzles Despite excavations, no human remains have been collected from these sites so it’s not possible to draw a firm conclusion about the efficacy of eDNA from this analysis alone. Meyer-Kaiser’s final report, which was shared with the DPAA in April, recommends excavation at three crash sites in the areas where the study found high abundance of human DNA sequences. However, it’s not clear whether that will happen. Excavation is necessary to establish the results as ground truths and show whether the patterns observed are actually associated with human remains, Meyer-Kaiser said. Without this step, “we don’t have a smoking gun,” Meyer-Kaiser added. Konsitzke agreed. “We need more insights to practice this. We need to correlate it with the recovery,” of human remains, he said. Further study is also needed to untangle one puzzling finding. Water samples from the two cold, deep test beds, in Saipan and Lake Huron, revealed a higher abundance of older human DNA sequences at sites without suspected remains than at those with suspected remains — an unexpected pattern given the former are located farthest from human activity. A similar pattern, however, was not observed for the two test beds with warmer water temperatures, in the shallower Saipan lagoon and in Italy. “In both the Saipan deep and the Great Lakes test bed, we had more human DNA at the control sites. That confused the hell out of me,” Meyer-Kaiser said. Perhaps, she suggested, the patterns they observed were the result of decades-old sewage runoff or even rain washing human DNA into the sea. “Human DNA is everywhere because humans are everywhere, and we’re constantly shedding,” Meyer-Kaiser noted. It’s possible, Konsitzke said, that the ocean is awash with human DNA, particularly in colder waters, where it might preserve for longer. The metagenomic analysis also revealed multiple species of microbes, most of which were not reliably associated with human remains, Meyer-Kaiser said. However, genes for a class of molecules called glycosaminoglycans that are common in animal skin, cartilage and blood vessels were detected at the site off Palermo — something she said she wants to investigate further. The ‘realm of possible’ It is important to apply new sciences to old questions such as those surrounding the search for lost troops, said Jennifer McKinnon, a maritime archaeologist at East Carolina University. McKinnon has researched the Avenger wreck and other World War II sites in Saipan both independently and with the DPAA, but she wasn’t involved in the pilot study. “I think it’s an exciting innovation in terms of trying to take new developing technologies and apply it to cold cases or sites where more information is needed. And if it turns out to have results that are going to benefit the mission of DPAA, I think it’s important to pursue,” she said of the eDNA findings. “It certainly would I think change the efficiency and efficacy of some of the missions if it proves to be useful.” What the DPAA will ultimately do with the eDNA findings reported by Meyer-Kaiser and her team, however, remains unclear. Meyer-Kaiser said she hopes in the future the DPAA will greenlight a phase 2 study and is looking for other wreck sites more definitively associated with human remains to give more credence to the approach. She also wants to study some of the microbes identified in the study in greater depth to see whether they can be reliably linked with animal bone and cartilage. But for now, identifying lost lives from the invisible traces they left behind remains out of reach. Stephen, the innovation chief at the DPAA, said the agency is going to take some time to “really discuss and evaluate” the research, adding that officials are also awaiting the results of another pilot study that makes use of imagery taken by underwater drones and machine learning. Stephen did express enthusiasm about the team’s work though, saying that he was particularly keen to further understand whether microbial DNA can be a reliable proxy for human DNA in these underwater environments. But he said he also didn’t want to raise the hopes of family members who might view eDNA as a way to quickly identify lost loved ones, noting that eDNA would be an investigatory tool to confirm the presence or absence of human DNA, not a way to identify a particular individual. “Generally, we were encouraged by the results, particularly the success that they had in extracting that degraded human DNA,” he said. “These are all stepping stones to be more informed and aware of what is in the realm of possible.”
