Former Cabanatuan POWs in makeshift hospital after the successful raid in January 1945. Credit: U.S. Army
On any given day between 1942 and 1943, between 5,000 and 8,000 Allied prisoners of war (POW) were held at the Cabanatuan prison camp by Japanese forces. By time the camp was liberated in January 1945—rescuing 489 POWs and 33 civilians—2,764 burials had occurred.
Anthropologists know that thanks to death records and cemetery layouts recovered just days after the successful raid.
“One of the things that made this camp different is personnel were allowed to keep track of who died,” said Suni Edson, assistant technical leader of the past accounting section at the Armed Forces DNA Identification Laboratory.
And while names and a map are a good starting point, the prison camp cemetery is still a mass grave.
“There are over 1,100 individual gravesites containing multiple individuals. All those who perished in a 24-hour period were believed to have been buried within a single grave. A single grave can have anywhere from 1 person to 79, which was found in the largest particular grave,” explained Edson during a presentation she gave at ISHI earlier this month.
Edson is part of the AFDIL team working to identify all POWs who died at the Cabanatuan Prison Camp, work that began in 2014 and is still ongoing. Although technically, one could say the research began all the way back in 1945.
Initial identification attempts
Soon after WWII ended in 1945, the American Graves Registry (AGRS) exhumed the Cabanatuan Prison Camp cemetery, preliminarily identifying 1,000+ POWs on the basis of identification tags or dental comparison. However, identifications based on ID tags come with a high margin of error, and dental identifications in 1945 don’t meet today’s standards. As a result, many of these individuals were misidentified.
During this initial period of processing, analysis, and identification—any subsequent attempts in the years to come—commingling of skeletal elements among individuals within the same common grave and between common graves likely occurred, in addition to the erroneous identifications.
In 1951, those yet to be identified were reburied at the Manila American Cemetery. Present-day analysis involves the disinterment and analysis of these remains.
Modern attempt
In September 2014, the AFDIL team started with Common Grave 717, as historical records indicated 14 buried individuals. But the team only found 10 individuals using mitochondrial DNA analysis.
“We did not expect this to be a challenge,” said Edson. “Our extraction technique had a 95% success regardless of element sample. We had maternal references for 12 of 14 missing individuals. But the remains were commingled to an extent we didn’t anticipate.”
Edson and the team couldn’t figure out why their DNA analysis was not working. (Keep in mind, in 2014, next-generation sequencing was not available to the extent it is today). Overall, the bone samples looked good, other than some indication of being submerged in water thanks to Manila’s 80 inches of rainfall per year.
Thanks to archival research, the team eventually discovered the remains were likely treated with a hardening compound that included 2% formalin. Formalin is “not a friend to DNA, as Edson says, since it binds into the histones of DNA and proteins in a skeletal matrix and prevents PCR amplification.
Thus, the researchers turned to “free DNA” and mitogenome next-generation sequencing (NGS).
“We had a 38% success rate originally, but were able to get to a 60% success rate by implementing NGS into the protocol,” said Edson.
The researchers were able to identify 18 mtDNA profiles. The grave was supposed to have 14 individuals, and the team successfully identified 12 of them.
“We probably had all 14 of the people were looking for, but now we see there is commingling between the graves themselves, which was not supposed to happen. The people in the grave were supposed to be the people in the grave. We also lacked two appropriate mitochondrial DNA references,” said Edson.
STRs to the rescue
Still, the identification of the individuals in Common Grave 717 was considered successful, enough so that the Defense POW/MIA Accounting Agency moved forward with disinterring caskets associated with other common graves.
“What we thought was going to be about 1,700 individuals, which is difficult in the first place, became the potential that none of those identifications done in 1950 were correct. So, we were looking at trying to identify all 2,768 people who were there,” said Edson.
For that level of identification, mitochondrial DNA would not be appropriate. The scientists needed STRs.
With a modified extraction protocol that used double the amount of skeletal material (.5 g to 1.0 g), the team conducted STR testing using a modified YSTR approach, as well as auto STR with the MiniFiler and PowerFlex Fusion.
To their surprise, it worked. The researchers recorded a success rate high of 23% for the modified YSTR testing.
“In this case, we think the formalin helped. Formalin binds to all proteins in your body, not jusy the histones. So, it bound into the bone and prevented the expansion of the osseous material during the water. When skeletal materials are soaked in water, they expand, a lot. It takes a while to remove the water from the bone so you can actually test the bone. In this instance, even though the bones were soaked in water since buried in 1950, they didn’t expand. We think that’s why the STR worked,” explained Edson.
Interestingly, the team noticed one particular common grave giving better results across all samples tested. The grave is not distinct in any way, and no preferential testing was performed during analysis. Still, Common Grave 225 is an outlier—which means more testing is needed to understand why.
“We’re going to examine the soil composition around the grave and see if there is something in particular that caused this to happen. Is there something going on that made it worse for everything else? Are there fertilizers or pesticides?,” said Edson.
The researchers hope they can figure out what the “x-factor” is and then perhaps develop a modified testing strategy to improve results.
“Because there are a lot left,” said Edson. “We’ve only disinterred 40 common graves, there are 1,100 left. Currently there are 218 mitotypes, which means we have a minimum of 218 people, but there is so much more left to do. If we can figure out anything that makes it easier, the better it will be.”
More recently, the scientists have conducted limited testing with some remains using SNP analysis. They have seen some success, but the work is still in validation.