by Charles Q. Choi, Contributor, Inside Science News. Originally published by Inside Science News. Read the original here.
A quick and accurate way to distinguish human blood from animal blood can prove key in crash investigations.
Now a new technique can rapidly tell the difference between human blood and that of nearly a dozen animal species, all without needing to destroy the samples in question. Such research could one day help police confirm or deny the human origins of bloodstains found at possible crime scenes.
Current blood exams usually destroy tested samples and any DNA evidence along with them. They can often also yield false positive results that falsely suggest a human presence.
In the new study, the scientists analyzed human and animal blood using a technique where they scraped dried blood samples onto a crystal and scanned them with infrared light. The infrared rays caused blood directly in contact with the crystal to emit a spectrum of a wide variety of wavelengths of light, which acted like a fingerprint unique to each kind of blood. This specific technique allows researchers to probe these samples without chemically altering the material in question.
The researchers trained their system on human, cat, dog, rabbit, horse, cow, pig, opossum and raccoon blood samples. The spectra the scientists collected from each sample appeared virtually identical to each other to the naked eye. However, advanced statistical analysis techniques spotted differences and helped correctly identified all but one of 290 spectra as either human or animal. (The lone error consisted of a raccoon sample mistaken as human.)
In addition, when samples from three other species -- deer, elk and ferret -- were included to further test the statistical model, it correctly classified all the samples. "That is encouraging -- it gives additional confidence that this model is doing what it is supposed to do," said biomedical engineer Ishan Barman at Johns Hopkins University, who did not take part in this research. "Often a model may perform very well with data very similar to what it is trained on, but there are open questions on how it might do with new kinds of data not used to construct the model."
The researchers noted the animals they chose are either common as pets, consumed by people, or can be involved in wildlife crimes or vehicle accidents.
For example, South Dakota's attorney general Jason Ravnsborg reported to police that he thought he hit a deer with his car on a Saturday night in September, but it was later found he had actually struck and killed a man.
"The attorney general Jason Ravnsborg case is a great example where our method could be extremely valuable in obtaining information on the blood origin and assessing whether a human or a deer was involved," said study lead author Ewelina Mistek-Morabito, a forensic chemist at the State University of New York at Albany. (The investigation into Ravnsborg may be completed by Christmas.)
The scientists also noted the kind of instrument they used for these analyses is sometimes portable, so this new technique could one day find use directly at crime scenes. "It's exciting to have new innovations to help us do even better with our evidence," said forensic scientist Ray Wickenheiser, director of the New York State Police Crime Laboratory System, who did not participate in this study.
In the future, the researchers hope to explore how to perform these analyses in the field, such as directly scanning bloodstains on the surfaces where they are found instead of having to scrape blood onto crystals. They also want to see how well these analyses work on bloodstains deposited on a variety of surfaces, exposed to different indoor and outdoor conditions, and contaminated with impurities, said study senior author Igor Lednev, a forensic chemist at SUNY Albany.
Mistek-Morabito and Lednev detailed their findings online Dec. 10 in the journal Communications Chemistry.
Photo credit: Alan Cleaver via Flickr