Implementing a biogeographic ancestry inference service for forensic casework

Abstract

The Centre of Forensic Sciences has validated the Precision ID Ancestry Panel on the Ion S5™ Massively Parallel Sequencing instrument for use in forensic casework. The focus of this paper is the development of reporting guidelines for implementation of the biogeographic ancestry inference service based on the Admixture Prediction results produced using the Torrent Suite™ Software (Thermo Fisher Scientific). The Admixture Prediction algorithm estimates the genetic ancestry of a sample using seven root populations (Europe, East Asia, Oceania, America, Africa, South Asia, and Southwest Asia). For individuals that declared a single ancestry, there was a high correlation between the declared ancestry and the ancestry predicted by the algorithm. However, some individuals with declared ancestries of Southern Europe, Southwest Asia, South Asia and Horn of Africa had Admixture Predictions that were composed of two or more root populations at 20% or greater. For individuals with known admixed ancestry, the major component of their declaration was included in their results in all but one case. Based on these results, reporting guidelines were developed and subsequently evaluated using the Admixture Predictions of additional samples. This paper discusses the development and evaluation of these reporting guidelines, along with an implementation plan for forensic casework.

Evaluation of the Ion AmpliSeq™ PhenoTrivium Panel: MPS-Based Assay for Ancestry and Phenotype Predictions Challenged by Casework Samples

“…Ancestry prediction was performed using the bootstrapping admixture analysis feature of Converge using the Precision ID Ancestry Panel Ancestry Frequency File v1.1. The frequency file contains genotype frequencies and population data for 146 SNPs of the Precision ID Ancestry panel and covers seven root populations created by hierarchal clustering of 66 populations from ALFRED based on allele frequencies [21,22]. As 163 of 165 of the Precision ID Ancestry SNPs were included in our panel, this corresponds to 145 SNPs (marked in yellow in Table S1) with available genotype frequencies and population data available for bootstrapping admixture analysis.…”

“…As 163 of 165 of the Precision ID Ancestry SNPs were included in our panel, this corresponds to 145 SNPs (marked in yellow in Table S1) with available genotype frequencies and population data available for bootstrapping admixture analysis. In the bootstrapping admixture analysis feature of Converge, admixture predictions are made based on a maximum likelihood approach used to predict the most likely admixture proportions across seven root populations (herein referred to as the core admixture algorithm): Africa (AFR), East Asia (EA), South Asia (SA), Southwest Asia (SWA), Europe (EU), America (AME), and Oceania (OCE) [21][22][23]. The predictions are bootstrapped across a random subset of sequenced SNPs, specified by the user in %, with each bootstrapping replication ran through the core admixture algorithm N times using a different subset of SNPs for each replication to capture uncertainty in the predictions.…”

“…The predictions are bootstrapped across a random subset of sequenced SNPs, specified by the user in %, with each bootstrapping replication ran through the core admixture algorithm N times using a different subset of SNPs for each replication to capture uncertainty in the predictions. The results are displayed as an average of the bootstrapping replications for each population group and a 95% confidence interval reflecting the probable range of variability of the estimated ethnicity percentages [21][22][23]. The predicted ancestry is presented as a percentage of each population with the corresponding likelihood.…”

Massive parallel sequencing in forensics: advantages, issues, technicalities, and prospects

“…ThemoFisher Scientific has designed and released a variety of Precision ID Panels for use on the Ion Torrent platform including those focused on STRs, identity SNPs, ancestry SNPs, and mitochondrial sequencing. The performance of the Ion Torrent instrument suite itself in respect of forensic applications has been tested extensively in the last 4 years for the genotyping of varied selected markers [35][36][37]. As relating to the currently available commercial panels, extensive validation studies by Pereira et al [35] and Al-Asfi et al [36] have been conducted of the 165 SNP Precision ID Ancestry Panel.…”

Forensic inference of biogeographical ancestry from genotype: The Genetic Ancestry Lab

“…The Centre for Forensic Sciences in Toronto, Canada, reports admixture from the seven root populations used by HID SNP Genotyper (Jin et al, ). They report four possible results: Single inclusion (one root population ≥80%, others ≤15%) Single mixed inclusion (one root population in the range 55–75%, others ≤15%) Double inclusion (two root populations ≥20%) Uninformative (at least three root populations ≥20%) …”