OsteoID

Overview

OsteoID is a forensic anthropology software program developed to aid law enforcement officers and crime scene personnel in the medico-legal investigation of recovered skeletal remains. Through uniform metrics of recovered samples, the software utilizes quantitative analyses to predict the likelihood of the bone being human in origin. OsteoID is currently under development through research supported by the United States Department of Justice and the National Institute of Justice, with the intention of being released as a free, web-based tool.

Development

Contributors

The OsteoID software development was proposed by Des Moines University Medical School Associate Professors of Anatomy Dr. Heather Garvin-Elling, a Forensic Anthropologist, and Dr. Rachel Dunn, a Vertebrate Paleontologist; in conjunction with Biological Anthropologist and Curator of Biological Anthropology at the National Museum of Natural History within the Smithsonian Institution system. Processing of the metrics, data, and photographs were aided by Noah Skantz, Merna Mohamed, and Nathan Kuttickat – Doctor of Osteopathic Medicine students at Des Moines University.

Funding

In 2018, the OsteoID proposal received grant funding of $185,453 USD from the National Institute of Justice under the solicitation of Research and Development in Forensic Science for Criminal Justice Purposes. This solicitation, which distributed a sum of $24,572,013 USD via 52 awards in 2018, required that funded projects meet at least one of two requirements:

1. "Increase the body of knowledge to guide and inform forensic science policy and practice, or..."
2. "Lead to the production of useful material(s), device(s), system(s), or method(s) that have potential for forensic application."

Sample Allocation

The development of OsteoID's preliminary software required samples of the skeletal remains of human and 27 non-human species (as demonstrated in Table 1). These species were selected based on their relationship to approximate human bone sizes, as well as species commonly cited in comparative osteology literature. These specimen samples, data, and photographs were obtained from the following institutions' collections:

• Smithsonian Institution / National Museum of Natural History - Washington, D.C.
• American Museum of Natural History - New York City, New York
• Des Moines University - Des Moines, Iowa
• University of California - Davis, California
• Mercyhurst University - Erie, Pennsylvania
• Washburn University - Topeka, Kansas

Table 1: Species Included in OsteoID Data

Mammals	Birds	Turtles
Human	Golden Eagle	Box Turtle
American Black Bear	Goose	Snapping Turtle
Brown Bear	Chicken
Cow	Mallard Duck
Elk	Turkey
Moose
White-Tailed Deer
Mule Deer
Pig
Horse
Sheep
Goat
Coyote
Wolf
Domestic Dog
Domestic Cat
Gray Fox
Red Fox
Raccoon
Opossum
Eastern Cotton-Tail Rabbit

Mechanism

Measurements

The foundation of the OsteoID software was developed from a database containing over 50,000 samples from the varying human and non-human species. Long bones – namely the humerus, radius, ulna, radio-ulna, femur, tibia, fibula, and fused metapodials – made up a majority of the data, as outlined in Table 2. The remainder of the measurements included skeletal measurements specific to certain species, such as the scapula, sacrum, and os coxae, amongst others.

Table 2: OsteoID Database Contents

Total Individuals / Animals	5,207
Total Bones Measured	18,867
Total Long Bones Measured	16,315
% Long Bones / Total Bones Measured	86.47%
Total Measurements Included	59,442
Total Long Bone Measurements	47,688
% of Long Bones / Total Measurements	80.23%

These measurements comprised of six basic metrics that formed the approach to measuring a recovered specimen's humeri, radii, ulnae, femora, or tibiae:

1. Maximum Length (MaxL)
2. Maximum Mediolateral Width of the Proximal Epiphysis (MaxPW)
3. Maximum Mediolateral Width of the Distal Epiphysis (MaxDW)
4. Maximum Anteroposterior Depth of the Distal Epiphysis (MaxDD)
5. Maximum Diameter of the Midshaft (MaxMidDD)
6. Minimum Diameter of the Midshaft (MinMidDD)

Calculations

In order to predict the likelihood that any given bone was human or non-human in origin, the data collected in the six basic metrics above were subjected to Linear Discriminant Analysis. Likewise, the normality of the data itself was analyzed via Kolmogorov-Smirnov Tests.

Software Efficiency

Efficiency

In December 2021, the OsteoID research team published their most recent findings in a Special Issue of the Journal of Biology dedicated to Recent Advances in Forensic Anthropological Methods and Research, which consisted of 11 published papers. The findings from that publication reported over a 98% accuracy in the software's ability to correctly identify a human bone. Likewise, the software demonstrated a >90% accuracy in correctly identifying a non-human bone as non-human. However, the software remained slightly less accurate in its ability to accurately predict which non-human species a given sample belongs to based on the preliminary measurements – correctly identifying the species with an accuracy between 77.7% - 89.1%.

In prior findings published in the American Journal of Physical Anthropology in March 2020, statistical analyses suggested that the humerus and femur were amongst the most reliable bones for the accurate identification of a non-human species. This abstract also primarily suggested that an increased number of measurements of a given specimen was positively correlated with an increased accuracy in identifying its origin.

Current Limitations

Several limitations to the OsteoID currently exist, most notably the limited number of species currently represented in the database. Expanding the scope of the database to include more species poses the potential to decrease the software's accuracy in correctly identifying the origin of a given sample. Likewise, the existence of variability within a given species (such as the vast array of sizes of the domestic dog) challenges the limitations of correcting for such variance via software calculations. Lastly, environmental considerations that may affect the bone morphology (e.g. burns, decay) have the potential to limit the software's efficacy in the setting of a forensics investigation.

Future Implications

In a Just Science podcast interview hosted by the Forensic Technology Center of Excellence (a program within the National Institute of Justice), OsteoID Co-Primary Investigator Dr. Heather Garvin identified that "approximately 30-40% of cases involving skeletal remains received by Forensic Anthropologists [such as herself] end up being animal bones;" the intention of OsteoID is to more effectively filter that percentage by providing crime scene personnel, death investigators, and law enforcement agents with easy-to-use decision making resources that don't require a background in anatomy. Garvin further cites the severe shortage of active professionals who have received licensure through the American Board of Forensic Anthropology, estimating that fewer than 90 board-certified forensic anthropologists are currently working in the United States.

References

This article "OsteoID" is from Wikipedia. The list of its authors can be seen in its historical and/or the page Edithistory:OsteoID. Articles copied from Draft Namespace on Wikipedia could be seen on the Draft Namespace of Wikipedia and not main one.

---