Evaluation of a custom QIAseq targeted DNA panel with 164 ancestry informative markers sequenced with the Illumina MiSeq

Abstract

Introduction of new methods requires meticulous evaluation before they can be applied to forensic genetic case work. Here, a custom QIAseq Targeted DNA panel with 164 ancestry informative markers was assessed using the MiSeq sequencing platform. Concordance, sensitivity, and the capability for analysis of mixtures were tested. The assay gave reproducible and nearly concordant results with an input of 10 and 2 ng DNA. Lower DNA input led to an increase in both locus and allele drop-outs, and a higher variation in heterozygote balance. Locus or allele drop-outs in the samples with less than 2 ng DNA input were not necessarily associated with the overall performance of a locus. Thus, the QIAseq assay will be difficult to implement in a forensic genetic setting where the sample material is often scarce and of poor quality. With equal or near equal mixture ratios, the mixture DNA profiles were easily identified by an increased number of imbalanced heterozygotes. For more skewed mixture ratios, the mixture DNA profiles were identified by an increased noise level. Lastly, individuals from Great Britain and the Middle East were investigated. The Middle Eastern individuals showed a greater affinity with South European populations compared to North European populations.

Figure 1

Sensitivity study. Percentage of correctly assigned genotypes, locus- and allele drop-outs for each dilution using a minimum read depth of 20 reads and a heterozygote balance of 0.3–3.0. (a) Samples analysed at UCPH, replicate 1; (b) samples analysed at UCPH, replicate 2; (c) samples analysed at KCL. Light grey indicates the % correctly assigned genotypes, medium grey indicates the % allele drop-outs, and dark grey indicates the % locus drop-outs.

Figure 2

Log transformed heterozygote balances in the sensitivity study. (a) UCPH. (b) KCL. Heterozygote balance was estimated as the number of reads for a nucleotide divided by the number of reads for the other nucleotide in the called genotype in the following order: A, C, G, and T. The outliers observed for 40 and 10 ng DNA input (a) represent the rs718505 locus as discussed in the sensitivity paragraph.

Figure 3

Mixture study. The read depth ratio for the nine loci for which the two individuals were opposite homozygotic. The read depth ratio was calculated as the number of reads for the allele of CEU1 divided by the number of reads for the allele of CEU2.

Figure 4

PCA plot of the studied populations and 1000 Genomes reference data. Meta-populations are listed in Table S2. ‘EASIA’ refers to East Asia and ‘SASIA’ refers to South-Central Asia.

Figure 5

STRUCTURE plot with K = 4 to K = 6 using 163 SNPs. Numbers above each population refers to the sample size. Population abbreviations on the horizontal axis: Gambia, Africa (GWD), Esan, Nigeria (ESN), Mende, Sierra Leone (MSL), Yoruba, Ibadan, Nigeria (YRI), Luhya, Kenya (LWK), African American SW USA (ASW), Puerto Rican, Puerto Rico (PUR), Colombian in Medellin, (CLM), Peruvians, Lima, Peru (PEL), Mexican, Los Angeles (MXL), Southern Han Chinese, China (CHS), Chinese Dai, Xishuangbanna (CDX), Kinh, Vietnam (KHV), Han Chinese, Beijing, China (CHB), Japanese, Tokyo, Japan (JPT), British, England and Scotland (GBR), Finnish, Finland (FIN), Iberians, Spain (IBS), Utah residents, North and West European ancestry (CEU), Toscans, Italy (TSI), British individuals, Great Britain (GBL) from this study, Syrians, Syria (SYR) from this study, Middle East (ME) from this study, Punjabi, Lahore, Pakistan (PJL), Bengali, Bangladesh (BEB), Tamil, Sri Lanka, from United Kingdom (STU), Teluga, India, from United Kingdom (ITU), and Gujarati, India, Houston Texas (GIH). Populations with a black line below the name are the populations genotyped in this study.