Forensic investigators move past old-school sleuthing techniques to collar criminals
Forensic biologist Silvana Tridico was puzzled by pubic hair.
Specifically, pubic hair samples donated by two volunteers.
She had just finished analyzing the bacteria stuck to the hair of seven people. If each hair sample carried unique mixes of bacteria, Tridico reasoned, investigators might have a new tool to help identify crime suspects. Hair bacteria, like fingerprints, could offer a forensic link between criminals and the bits of bodily debris they left behind.
But two of the hair samples held nearly identical microbe populations. “I thought I’d made a mistake,” Tridico says. She repeated the analysis, and still, the hairs’ bacteria matched. One explanation came to mind.
“I said to my partner, ‘I think they’ve had sex.’ ”
Tridico was right. Two of the study’s participants had, in fact, had sex 18 hours before snipping off strands of their hair for Tridico’s analysis. Their bacteria apparently mingled so much, she says, that their microbial medleys became indistinguishable. Since the telltale traces lingered for so many hours — even after the volunteers had showered — Tridico thinks the technique has the potential to match sex offenders to their victims.
“I’m so pumped,” says Tridico, of Murdoch University in Perth, Australia. “I really think it’s got traction.”
Though still in early stages, her technique and other advances signal the rumblings of a seismic shift in the forensics field. One day, tiny microbes could hand investigators big clues. Bacteria shed from people’s hair, skin and footprints, or fungi hidden in specks of dust, could help place suspects at the scene of a crime. And just a whiff of odor clinging to a fingernail or seeping from a dead body could help investigators track a missing person — or corpse.
As new techniques gain their scientific footing, researchers are also shoring up classic forensic tools that have taken recent high-publicity hits. Some of the most time-honored techniques, such as fingerprint analysis, have been resting on rickety foundations. In the last few years, researchers have taken a closer look at forensic science’s tarnished old hide. Bit by bit, they’re tugging it into the 21st century.
Story continues below slideshow
Traces of trouble
If justice is a woman clutching sword and scales, then forensic science is a sprawling beast, with a hodgepodge of tools stuffed in its fists.
These tools target almost every mark a person can leave behind: fingerprints, blood spatters, handwriting, DNA, hair and more. Each trace represents a little shred of history that might help investigators piece together a picture of the past.
But not all forensic science disciplines are created equal. Some, such as fingerprint analysis, are steeped in more than a century of tradition but not a lot of data. Seasoned experts train apprentices in the craft, and examiners’ opinions can carry too much weight.
Criminal forensics’ shortcomings drew international attention in 2004, when the FBI bungled the investigation of the Madrid train bombings. The bureau’s lab matched a single smudged fingerprint on a plastic bag to an Oregon attorney named Brandon Mayfield — an innocent man. Mayfield spent two weeks in jail before the Spanish National Police identified an Algerian man as the fingerprint’s actual source. The U.S. government formally apologized to Mayfield in 2006 and agreed to pay him $2 million.
The misidentification of Mayfield sparked renewed scrutiny of forensic techniques, says John Butler, a forensic DNA scientist at the National Institute of Standards and Technology in Gaithersburg, Md. “It really woke people up.”
Soon after, a National Academies panel poked into forensics’ scientific nooks and crannies. In 2009 it released a damning report. The story was grim: Deep cracks ran through several forensic bedrocks — especially those based on expert interpretations of patterns.
Experts often reach conclusions that are just too black or white. Testimony that two fingerprints “match,” for example, implies that an examiner’s findings are definitive.
But that’s never the case, says Christophe Champod, a forensic scientist at the University of Lausanne in Switzerland, who reviewed recent advances in fingerprint identification in the August 5 Philosophical Transactions of the Royal Society B.
“If the claim is that you’re able to exclude everyone on Earth but the person of interest,” he says, “that is ridiculous.”
Fingerprints and other forensic analyses can offer only probabilities, Butler says. But no one has a good handle on how to calculate them. Unlike DNA analyses, fingerprints lack the population data that would help pin down how certain an examiner is of a match.
The report exposed problems in other areas too. There’s little evidence that handwriting and bite-mark analyses can reliably identify people. Blood spatter experts read too much into stain patterns.
“It was a very hard-hitting report,” says study director Anne-Marie Mazza of the National Academies’ Committee on Science, Technology and Law in Washington, D.C. “It said that the system was broken,” she says, “that there was a lack of solid scientific basis for a number of techniques.”
The report didn’t shock everyone, Champod says. But it dropped a bomb on the public and members of the judiciary. “They were living with the illusion that all the science had been done for many, many years,” he says.
In reality, it hadn’t. Some disciplines, such as DNA and chemical analyses, were considered scientifically sound. But many techniques needed more grounding. So the National Academies report proposed a roadmap to improvement. In the six years since the report’s publication, scientists have begun dreaming up new forensic tools to replace or work alongside those that are less than reliable. They’re also beefing up the old standards.
“There’s been a bit of a paradigm shift,” Mazza says. “But change takes time. New ways of thinking take time.”
Americans may have adopted a new way of thinking about one forensic discipline this year, after reports of a more recent FBI fumble. The problem this time: microscopic hair analysis.
In April, the Justice Department and the FBI released a statement exposing errors made by FBI hair examiners. The magnitude of the misstep was breathtaking. In cases where microscopic hair comparisons linked a defendant to a crime, examiners made mistakes 96 percent of the time. In 33 out of 257 of these flawed cases, defendants received the death penalty.
Traditionally, examiners scrutinize hairs under a microscope, marking off distinctive features, such as shape and pigment patterns, inch by careful inch. If a strand of hair from a suspect looks like one found at a crime scene, prosecutors can tie a suspect to a crime.
But no one knows just how widespread different hair features are. So there’s no way to figure out the odds that one suspicious strand of hair actually belongs to a particular suspect. What’s more, microscopic hair analysis depends on a skillful eye. “It’s totally based on your expertise and abilities,” Tridico says. Overstate the evidence, and innocent people could go to jail, she says.
And that could happen, the investigation noted.
“The FBI had people who were saying things that were not supportable,” says forensic geneticist Bruce Budowle, who worked at the FBI Laboratory Division for decades. The National Academies report in 2009 had uncovered a similar problem. In court, some hair examiners would state the odds that a hair belonged to a certain suspect, despite the fact that no statistical data existed to back up the claim. No one had even hammered out what it meant to say that two hairs “matched.”
“I really hate that term, ‘match,’ ” says Tridico, an expert at analyzing human and animal hairs. A jury might think it means there’s no doubt a hair came from a certain suspect, she says. But there’s always doubt. An examiner could potentially use hair analysis to narrow a pool of suspects, but it’s just not possible to use the technique to identify an individual.
Today, most labs have moved away from microscopic hair comparisons. Labs rely more on a DNA-based technique called mitochondrial analysis to tell two hairs apart. In the technique, examiners collect DNA from within the tiny mitochondria that power a cell. These organelles are stuffed with their own DNA, which examiners can analyze for patterns and use to link hair samples from crime scenes to suspects.
Microbes are the new fingerprints
Tridico thinks hair analysis might improve with a metagenomics approach — going beyond the victim’s or the suspect’s DNA and instead looking at the bacteria loaded on a particular strand of hair, for example. Microbes on pubic hair differ from person to person, Tridico and colleagues reported last year in Investigative Genetics.
Source: S. Tridico et al/Investigative Genetics 2014
What’s more, pubic hair bacteria could reveal more than a person’s identity: They could even offer hints about lifestyle. One sample Tridico tested was covered in marine bacteria. “I thought it was really bizarre,” she says. Then she found out that the hair belonged to someone who swam in the ocean every day.
Tridico thinks that hair bacteria could one day help investigators glean other telling details about suspects, including their ethnicity (different ethnic groups can carry different assortments of bacteria in their mouths and guts, so hair bacteria may differ too). That day is still a long way off, Tridico cautions. Big hurdles stand in the way of making new advances that would be accepted in the courtroom.
New forensic techniques face the same troubles vexing old ones. They need to be scientifically validated, they need to be reliable and, to become commonplace in crime labs, they need to be simple and cheap. All this requires money for research, which has been lacking.
Still, the idea of microbes as an identification tool is catching on.
Skin bacteria lingering on a computer keyboard can be matched to the person who touched it, scientists reported in 2010. “We’re constantly leaving skin bacteria behind,” says study coauthor Noah Fierer, a microbial ecologist at the University of Colorado Boulder. So if investigators can’t scrape up enough human DNA to test, they could potentially analyze bacterial traces instead.
Still, like Tridico, Fierer thinks microbe-based identification needs more testing. Scientists need to confirm how well it works outside the lab. And they’ll need to gather a lot more data from a lot more people to find out whether a person’s microbial signature is unique. “We’re not locking people up in jail with this method,” he says. “It’s just a proof of concept.”
In June, researchers backed up the concept with some population data. Looking at human gut bacteria from poop, researchers calculated that they could pick out individuals from a pool of hundreds of people. They reported their results in the Proceedings of the National Academy of Sciences.
The DNA of bacteria and other tiny organisms could even serve as a kind of biological tracking device. Bacteria from the floor, for example, can transfer to shoes, leaving an invisible record of a person’s travels, ecologists suggested in May in Microbiome. And fungal spores riding on specks of dust can reveal roughly where in the United States those dust specks came from, Fierer and colleagues reported in April in PLOS ONE.
Fierer’s team analyzed DNA from traces of fungi in 928 dust samples from across the country, and mapped where different fungus species tended to cluster geographically. The map let the researchers pinpoint a dust sample’s origin within about 230 kilometers. Given the roughly 4,500 kilometers that stretch between east and west coasts, Fierer’s technique could help investigators infer a person’s general location.
“Let’s say someone sent you a package with a bomb in it,” he says. “The idea is that we could look at the dust in the sample and identify where the package had been shipped from.”
Scent of a human
In some cases, the barest traces of an odor — a few chemicals wafting in the air — may implicate a perp. Living or dead, people may leave behind a telltale scent.
The scent of human decomposition was first admitted into a U.S. court of law in 2011, at the trial of Casey Anthony. After a utility worker discovered the skeletal remains of her 2-year-old daughter, Caylee, prosecutors accused Anthony of first-degree murder. Odor evidence hinted that a dead body had spent time in the trunk of Anthony’s car: Two cadaver dogs had picked up the scent, as did a chemical analysis. In court, a forensic scientist reported that compounds in the air matched those known to leak from decaying bodies.
But no one has spelled out a reliable chemical signature for the scent of death, says analytical chemist Kenneth Furton of Florida International University in Miami.
“I testified that I didn’t believe the science was sufficient,” says Furton, who studies the odor of humans and their remains.
The odor and other forensic evidence presented in the case weren’t enough to convict. On July 5, 2011, a jury found Anthony not guilty of murder. She was released two weeks later, having already served three years for lying to police.
The scent of death is a slippery thing. “Once you die,” Furton says, “the chemicals you give off start to change almost immediately — and they vary over time.”
Within minutes of death, skin begins to blister and bag, and cells dissolve. A few days later, when microbes digest soft tissue, the body bloats and takes on a greenish hue. When the skin splits and slips away from the body, gases seep out, and muscles and fat start to decay. Bones are the last to crumble (SN Online: 7/22/15).
Each stage of the grisly process releases a different brew of chemicals. Identifying the ones that make up death’s signature scent is like throwing a dart at a clock and trying to hit the second hand. Still, scientists hoping to pin down the scent of death have seen a few promising glimmers.
“I think it can be done,” Furton says. Last year, he and colleagues picked out odors that could discern the stages of mouse decomposition, from fresh to advanced decay. Also in 2014, Belgian researchers writing in Analytical and Bioanalytical Chemistry suggested odor-trapping and analysis methods that could help scientists find chemicals specific to human corpses. Had scientists already identified these chemicals, the scent clinging to Anthony’s trunk might have been more telling.
Though the scent of death remains elusive, scientists have begun using odors to identify the living. Just the chemicals wafting off people’s bodies can distinguish them from others, Furton’s lab reported in 2013 in Forensic Science International.
Furton’s team collected nail clippings, hair snippets, saliva and cotton gauze pressed between the palms of 20 volunteers. Then the researchers analyzed odor compounds hovering in the air above the samples with a mass spectrometer.
The soap people use and food they eat contribute to their odor. After whittling those away to uncover the genetic component of scent, Furton’s team found that each person’s samples gave off a distinguishing chemical blend. “It’s almost like a human bar code,” Furton says.
He thinks that scent bar codes could offer investigators a clue when other biological evidence is scarce. Even if a perpetrator left a crime scene wiped clean of fingerprints or DNA, Furton says, “it’s very possible they would still leave behind human scent.”
And unlike some types of trace evidence, such as fingerprints, scent wouldn’t rely so much on examiners’ opinions.
Slow and steady
Matching a clean fingerprint with one inked on an official, 10-print card can be easy. Examiners run a print through a computer program that scans millions of other prints for similar patterns, and — voilà! — the program spits out a match.
The microbes living on people’s hair and skin could transfer to anything they touch (colored remnants shown in video). Those invisible traces may help investigators identify or track a suspect. BioBE Center/Vimeo
But analyzing the smudgy, dirty, partial prints picked up from crime scenes requires more finesse. Computer programs aren’t as good as the human eye at spotting details in low-quality prints. That’s where the experts come in.
Fingerprint examiners look for distinguishing features — the forked ridge of a whorl or the end of an arch, for example — and manually punch the details into a print-scanning program. But these details rely on subjective judgments, and the reliability of these judgments can vary. “I always worry when people say, ‘I’ve been doing this for 30 years, therefore I know what I’m doing,’ ” says Budowle, now at the University of North Texas Health Science Center in Fort Worth.
Rather than dumping fingerprints as an investigative tool, scientists have begun shoring up their scientific foundation (SN: 5/31/14, p. 14). In the Proceedings of the National Academy of Sciences in July, researchers studying 12 years of police fingerprint records confirmed what examiners had believed since the late 1800s: Fingerprints don’t change much over time. So a long-term print database makes sense.
Meanwhile, researchers have been trying to better understand and improve the technique’s reliability. Last year, Champod’s team reported in the Journal of Forensic Sciences that analyzing the itty-bitty pores dotting people’s fingertips could improve the reliability of fingerprint analysis. The metric might give examiners another way to differentiate between suspects.
These advances mark some of the forensic field’s steps toward greater scientific credibility. “We have a better handle on the problems than we did six years ago when the National Academies report came out,” Budowle says.
Forensic scientists are starting to fortify weak spots, but progress will be gradual, Butler cautions.
“You have to make changes as you go along,” he says. “You can’t shut down the entire criminal justice system and wait till we get everything perfect.”
This story appears in the September 5, 2015, Science News with the headline, "Wanted: Crime-solving bacteria and body odor: New techniques may boost credibility of forensic science."
Yoon, S., and A.K. Jain. Longitudinal study of fingerprint recognition. Proceedings of the National Academy of Sciences. Vol. 112, July 14, 2015, p. 8555. doi: 10.1072/pnas.1410272112.
N.S. Grantham et al. Fungi identify the geographic origin of dust samples. PLOS ONE. Published online April 13, 2015. doi: 101371/journal.pone.0122605.
S.R. Tridico et al. Metagenomic analysis of bacteria on human hairs: a qualitative assessment for applications in forensic science. Investigative Genetics. Published online December 16, 2014. doi: 10.1186/s13323-014-0016-5.
M.Sc., A.A., and C. Champod. Integration of pore features into the evaluation of fingerprint evidence. Journal of Forensic Sciences. Vol. 59, January 2014, p. 82. doi: 10.1111/1556-4029.12302
J.S. Brown et al. Applicability of emanating volatile organic compounds from various forensic specimens for individual differentiation. Forensic Science International. Vol. 226, March 10, 2013, p. 173. doi: 10.1016/j.forsciint.2013.01.008
N. Fierer et al. Forensic identification using skin bacterial communities. Proceedings of the National Academy of Sciences. Vol. 107, April 6, 2010, p. 14. doi: 10.1073/pnas.1000162107
E.A. Franzosa et al. Identifying personal microbiomes using metagenomic codes. Proceedings of the National Academy of Sciences. Vol. 112, June 2, 2015, p. E2930. doi:10.1073/pnas.1423854112.
E. Rosier et al. Development and validation of a new TD-GC/MS method and its applicability in the search for human and animal decomposition products. Analytical and Bioanalytical Chemistry. Vol. 406, March 16, 2014, p. 3611. doi: 10.1007/s00216-014-7741-8.