High-performance deep learning pipeline predicts individuals in mixtures of DNA using sequencing data.

In this study, we proposed a deep learning (DL) model for classifying individuals from mixtures of DNA samples using 27 short tandem repeats and 94 single nucleotide polymorphisms obtained through massively parallel sequencing protocol. The model was trained/tested/validated with sequenced data from 6 individuals and then evaluated using mixtures from forensic DNA samples. The model successfully identified both the major and the minor contributors with 100% accuracy for 90 DNA mixtures, that were manually prepared by mixing sequence reads of 3 individuals at different ratios. Furthermore, the model identified 100% of the major contributors and 50-80% of the minor contributors in 20 two-sample external-mixed-samples at ratios of 1:39 and 1:9, respectively. To further demonstrate the versatility and applicability of the pipeline, we tested it on whole exome sequence data to classify subtypes of 20 breast cancer patients and achieved an area under curve of 0.85. Overall, we present, for the first time, a complete pipeline, including sequencing data processing steps and DL steps, that is applicable across different NGS platforms. We also introduced a sliding window approach, to overcome the sequence length variation problem of sequencing data, and demonstrate that it improves the model performance dramatically.

Massively parallel sequencing analysis of nondegraded and degraded DNA mixtures using the ForenSeq™ system in combination with EuroForMix software.

Massively parallel sequencing (MPS) technologies enable the simultaneous analysis of short tandem repeats (STRs) and single nucleotide polymorphisms (SNPs). MPS also enables the detection of alleles of the minor contributors in imbalanced DNA mixtures. In this study, 59 STRs (amelogenin, 27 autosomal STRs, 7 X-STRs, and 24 Y-STRs) and 94 identity-informative SNPs of 119 unrelated Taiwanese (50 men, 69 women) were sequenced using a commercial MPS kit. Forty-eight nondegraded and 44 highly degraded two-person artificial DNA mixtures with various minor to major ratios (1:9, 1:19, 1:29, 1:39, 1:79, and 1:99) were analyzed to examine the performance of this system for detecting the alleles of the minor contributors in DNA mixtures. Likelihood ratios based on continuous model were calculated using the EuroForMix for DNA mixture interpretation. The STR and SNP genotypes of these 119 Taiwanese were obtained. Several sequence variants of STRs were observed. Using EuroForMix software based on the sequence data of autosomal STRs and autosomal SNPs, 97.9% (47/48) and 97.7% (42/43) of minor donors were accurately inferred among the successfully analyzed nondegraded and degraded DNA mixtures, respectively. In conclusion, combined with EuroForMix software, this commercial kit is effective for assignment of the minor contributors in nondegraded and degraded DNA mixtures.

A single nucleotide polymorphism panel for individual identification and ancestry assignment in Caucasians and four East and Southeast Asian populations using a machine learning classifier.

Single nucleotide polymorphism (SNP) profiling is an effective means of individual identification and ancestry inferences in forensic genetics. This study established a SNP panel for the simultaneous individual identification and ancestry assignment of Caucasian and four East and Southeast Asian populations. We analyzed 220 SNPs (125 autosomal, 17 X-chromosomal, 30 Y-chromosomal, and 48 mitochondrial SNPs) of the DNA samples from 563 unrelated individuals of five populations (89 Caucasian, 234 Taiwanese Han, 90 Filipino, 79 Indonesian and 71 Vietnamese) and 18 degraded DNA samples. Informativeness for assignment (In) was used to select ancestry informative SNPs (AISNPs). A machine learning classifier, support vector machine (SVM), was used for ancestry assignment. Of the 220 SNPs, 62 were individual identification SNPs (IISNPs) (51 autosomal and 11 X-chromosomal SNPs) and 191 were AISNPs (100 autosomal, 13 X-chromosomal, 30 Y-chromosomal, and 48 mitochondrial SNPs). The 51 autosomal IISNPs offered cumulative random match probabilities (cRMPs) ranging from 1.56 × 10-21 to 3.16 × 10-22 among these five populations. Using AISNPs with the SVM, the overall accuracy rate of ancestry inference achieved in the testing dataset between Caucasian, Taiwanese Han, and Filipino populations was 88.9%, whereas it was 70.0% between Caucasians and each of the four East and Southeast Asian populations. For the 18 degraded DNA samples with incomplete profiling, the accuracy rate of ancestry assignment was 94.4%. We have developed a 220-SNP panel for simultaneous individual identification and ethnic origin differentiation between Caucasian and the four East and Southeast Asian populations. This SNP panel may assist with DNA analysis of forensic casework.

A panel of 130 autosomal single-nucleotide polymorphisms for ancestry assignment in five Asian populations and in Caucasians.

Ancestry informative single-nucleotide polymorphism (AISNP) panels for differentiating between East and Southeast Asian populations are scarce. This study aimed to identify AISNPs for ancestry assignment of five East and Southeast Asian populations, and Caucasians. We analyzed 145 autosomal SNPs of the 627 DNA samples from individuals of six populations (234 Taiwanese Han, 91 Filipinos, 79 Indonesians, 60 Thais, 71 Vietnamese, and 92 Caucasians) using arrays. The multiple logistic regression model and a multi-tier approach were used for ancestry classification. We observed that 130 AISNPs were effective for classifying the ethnic origins with fair accuracy. Among the 130 AISNPs, 122 were useful for stratification between these five Asian populations and 64 were effective for differentiating between Caucasians and these Asian populations. For differentiation between Caucasians and Asians, an accuracy rate of 100% was achieved in these 627 subjects with 50 optimal AISNPs among the 64 effective SNPs. For classification of the five Asian populations, the accuracy rates of ancestry inference using 20 to 57 SNPs for each of the two Asian populations ranged from 74.1% to 100%. Another 14 degraded DNA samples with incomplete profiling were analyzed, and the ancestry of 12 (85.7%) of those subjects was accurately assigned. We developed a 130-AISNP panel for ethnic origin differentiation between the five East and Southeast Asian populations and Caucasians. This AISNP set may be helpful for individual ancestral assignment of these populations in forensic casework.

Genotyping of 75 SNPs using arrays for individual identification in five population groups.

Single nucleotide polymorphism (SNP) typing offers promise to forensic genetics. Various strategies and panels for analyzing SNP markers for individual identification have been published. However, the best panels with fewer identity SNPs for all major population groups are still under discussion. This study aimed to find more autosomal SNPs with high heterozygosity for individual identification among Asian populations. Ninety-six autosomal SNPs of 502 DNA samples from unrelated individuals of five population groups (208 Taiwanese Han, 83 Filipinos, 62 Thais, 69 Indonesians, and 80 individuals with European, Near Eastern, or South Asian ancestry) were analyzed using arrays in an initial screening, and 75 SNPs (group A, 46 newly selected SNPs; groups B, 29 SNPs based on a previous SNP panel) were selected for further statistical analyses. Some SNPs with high heterozygosity from Asian populations were identified. The combined random match probability of the best 40 and 45 SNPs was between 3.16 × 10(-17) and 7.75 × 10(-17) and between 2.33 × 10(-19) and 7.00 × 10(-19), respectively, in all five populations. These loci offer comparable power to short tandem repeats (STRs) for routine forensic profiling. In this study, we demonstrated the population genetic characteristics and forensic parameters of 75 SNPs with high heterozygosity from five population groups. This SNPs panel can provide valuable genotypic information and can be helpful in forensic casework for individual identification among these populations.

Analysis of nondegraded and degraded DNA mixtures of close relatives using massively parallel sequencing.

Identification of the minor contributor in DNA mixture of close relatives remains a dilemma in forensic genetics. Massively parallel sequencing (MPS) can analyze multiple short tandem repeats (STRs) and single nucleotide polymorphism (SNPs) concurrently and detect non-overlapping alleles of the minor contributors in DNA mixtures. A commercial kit for MPS of 59 identity informative STRs (iiSTRs) and 94 autosomal identity-informative SNPs (iiSNPs) was used to analyzed 34 nondegraded and 33 highly degraded two-person artificial DNA mixtures of close relatives with various minor to major ratios (1:9, 1:19, 1:29, 1:39, 1:79, 1:99). EuroForMix software was used to determine the minor contributors in the mixtures based on the likelihood ratios calculated from the MPS data, and relMix software was used to perform kinship analysis of the contributors. The STRs and SNPs of the 34 nondegraded and 33 degraded DNA mixtures were genotyped using MPS. Using EuroForMix based on the genotypes of autosomal iiSTRs and autosomal iiSNPs, 82.4% (28/34) and 54.5% (18/33) of minor donors could be accurately assigned for the nondegraded and degraded DNA mixtures, respectively. The relMix software correctly inferred the relationship between contributors in 97.1% (33/34) of nondegraded mixtures and in 97.0% (32/33) of degraded mixtures. In conclusion, combined EuroForMix and MPS data of STRs and SNPs can assist in the assignment of minor donors in nondegraded DNA mixtures of close relatives, and relMix can be used to infer relationship among contributors.

A 1204-single nucleotide polymorphism and insertion-deletion polymorphism panel for massively parallel sequencing analysis of DNA mixtures.

Massively parallel sequencing (MPS) technology enables the simultaneous analysis of a huge number of single nucleotide polymorphisms (SNPs) and insertion-deletion polymorphisms (indels). MPS also enables the detection of the alleles of minor contributors in a highly unbalanced DNA mixture. In this study, we established a 1204-marker panel optimized for MPS consisting of 987 autosomal markers (964 SNPs and 23 indels), 27 X-chromosome SNPs, 61 Y-chromosome markers (56 SNPs and 5 indels), and 129 mitochondrial SNPs. The DNA samples of six unrelated individuals (two men and four women), 26 nondegraded DNA mixtures (with minor to major ratios of 1:29, 1:39, 1:79, and 1:99), and eight highly artificially degraded DNA mixtures (with minor to major ratios of 1:29, 1:39, 1:79, and 1:99) were analyzed through MPS by using the panel. A scoring system was developed to determine the minor contributors in DNA mixtures based on the genotypes identified using MPS. The genotypes of the 1204 markers were successfully profiled through MPS by using the custom-designed panel. The efficiency of MPS for analyzing these highly degraded samples was lower than that for analyzing nondegraded samples. All minor contributors in the 26 nondegraded and 8 degraded DNA mixtures were accurately assigned using this scoring system based on 964 autosomal SNPs. An association between the observed reads ratio and theoretical ratio of the minor component was noted for nondegraded mixtures. In conclusion, we established a 1204-marker individual identification panel for MPS that successfully analyzed autosomal, X-chromosome, Y-chromosome, and mitochondrial SNPs and indels simultaneously. In combination with the newly developed scoring system, the panel can accurately identify minor contributors in nondegraded and highly degraded DNA mixtures.