Macromorphoscopic trait expression in a cranial sample from Medellín, Colombia

Area of Science:

• Forensic Anthropology
• Bioarchaeology
• Population Genetics

Background:

• Human identification methods in forensic anthropology are often population-specific.
• Adapting these methods requires contemporary data and assessment of within-group variation.
• Colombia's complex demographic history complicates ancestry assessment.

Purpose of the Study:

• To document cranial macromorphoscopic trait variation in a Colombian sample.
• To assess the applicability of US-developed forensic methods to Colombian populations.
• To explore the potential for finer-level ancestry identification within Colombia.

Main Methods:

• Analysis of cranial macromorphoscopic traits in 244 individuals from Antioquia, Colombia.
• Utilizing artificial neural networks (ANNs) for classification based on birthplace.
• Comparing ANNs' performance on pooled Colombian data against modern American samples.

Main Results:

• ANNs showed low initial classification accuracy for birthplace within Colombia (18.6% validation).
• Overfitting was observed in the initial ANN models.
• When pooled with American data, the best ANN model achieved 48.4% validation accuracy.

Conclusions:

• Macromorphoscopic trait data alone is insufficient for fine-grained regional ancestry identification in Colombia.
• Existing forensic anthropological methods require adaptation and more representative data for diverse populations.
• Further research is needed to improve ancestry assessment accuracy in complex populations.