Analysis of the Accuracy of AncesTrees Software in Ancestry Estimation in Brazilian Identified Sample

Larissa Chaves Cardoso Fernandes,David Navega,Evelyne Pessoa Soriano,Maria Izabel Cardoso Bento,Eduardo Daruge Júnior,Patrícia Moreira Rabello,Eugénia Cunha

doi:10.4236/aa.2021.112011

Larissa Chaves Cardoso Fernandes, David Navega + Show 5 more

Open Access

https://doi.org/10.4236/aa.2021.112011

Copy DOI

Abstract

In the present study a software tool for craniometric ancestry estimation, AncesTrees, was evaluated in an identified Brazilian skeletal sample with known self-reported ancestry. Twenty-three craniometric measures were obtained from each skull and analyzed using AncesTrees software, with two classification strategies—tournamentForest and ancestralForest algorithm. The tournamentForest (53.54%) and ancestralForest algorithms with three ancestry groups (50.96%) were more accurate to classify Europeans, while the ancestralForest algorithm with six (50.00%) and two (67.64%) groups were more accurate to estimate the ancestry of African descents. Admixed ancestry specimens were classified predominantly as European descent. The use of the ancestralForest algorithm considering only European and African origin (58.42%) was the most accurate setup for ancestry estimation in Brazilian skulls. Supervised classification algorithms and tools such as the AncesTrees work based on data analysis and pattern matching, and there is no Brazilian sample in its database, the software showed a low accuracy Brazilian samples. The incorporation of representative craniometric data obtained from Brazilian skulls into the software database may significantly increase the accuracy of ancestry estimates.

Highlights

Several programs and software tools have been developed in recent years to tackle the challenging task of ancestry estimation from skeletal remains in forensic anthropology (FA)
The tournamentForest (53.54%) and ancestralForest algorithms with three ancestry groups (50.96%) were more accurate to classify Europeans, while the ancestralForest algorithm with six (50.00%) and two (67.64%) groups were more accurate to estimate the ancestry of African descents
Supervised classification algorithms and tools such as the AncesTrees work based on data analysis and pattern matching, and there is no Brazilian sample in its database, the software showed a low accuracy Brazilian samples

Summary

Introduction

Several programs and software tools have been developed in recent years to tackle the challenging task of ancestry estimation from skeletal remains in forensic anthropology (FA) These computational tools make use of statistical and machine learning algorithms to solve a mathematical problem that abstractly speaking involves allocate objects to predefined classes the name classifiers or classification algorithms. Some prime examples of such tools are FORDISC (Ousley & Jantz, 2013), CRANID (Wright, 1992), COLIPR (Urbanová & Králík, 2008), 3D-ID (Slice & Ross, 2009) and AncesTrees (Navega et al, 2015) The latter, which will be the focus of the present study, was developed in 2015 by Portuguese researchers to quantitatively estimate ancestry based on 23 craniometric measures. This tool classifies the human skull using the random forest algorithm (Breiman, 2001), a non-linear and non-parametric ensemble-based classification technique that uses hundreds to thousands of classification decision trees as base models

Methods

Results

Discussion

Conclusion