RESUMO
The objective of the investigation was to test a technique originating in the United States to verify the accuracy of the ancestry estimate in a sample of Brazilian skulls. The sample consisted of 120 skulls of Brazilian adults from the collection of the Institute of Education and Research in Forensic Sciences (IEPCF), Guarulhos-SP, Brazil, with ancestry determined by routine examination. In the skull, the following structures were analyzed: anterior nasal spine (ANS), inferior nasal aperture (INA), interorbital breadth (IOB) nasal aperture width (NAW), nasal bone structure (NBS), and post bregmatic depression (PBD). The Hefner and Ousley technique was applied to classify ancestry. Data were registered and correctness percentage was obtained. To improve ancestry prediction, Machine Learning techniques were used. Regarding the error of the studied method and the correctness of the complete sample, the study presented a percentage above 50%, with values such as 56.33% for whites and 53.07% for non-whites. The most frequent ANS parameter was marked (37.5%), INA straight (31.7%), IOB wide (38.3%), NAW medium (32.5%) NBS triangular (29.2%), and PBD present (78.3%). The application of the method in the sample of the study showed a limitation to estimate ancestry, suggesting a need for adaptation for its applicability in the Brazilian population.