Eye color and the prediction of complex phenotypes from genotypes

Exploring Eye, Hair, and Skin Pigmentation in a Spanish Population: Insights from Hirisplex-S Predictions

BACKGROUND/OBJECTIVES

Understanding and predicting human pigmentation traits is crucial for individual identification. Genome-wide association studies have revealed numerous pigmentation-associated SNPs, indicating genetic overlap among pigmentation traits and offering the potential to develop predictive models without the need for analyzing large numbers of SNPs.

METHODS

In this study, we assessed the performance of the HIrisPlex-S system, which predicts eye, hair, and skin color, on 412 individuals from the Spanish population. Model performance was calculated using metrics including accuracy, area under the curve, sensitivity, specificity, and positive and negative predictive value.

RESULTS

Our results showed high prediction accuracies (70% to 97%) for blue and brown eyes, brown hair, and intermediate skin. However, challenges arose with the remaining categories. The model had difficulty distinguishing between intermediate eye colors and similar shades of hair and exhibited a significant percentage of individuals with incorrectly predicted dark and pale skin, emphasizing the importance of careful interpretation of final predictions. Future studies considering quantitative pigmentation may achieve more accurate predictions by not relying on categories. Furthermore, our findings suggested that not all previously established SNPs showed a significant association with pigmentation in our population. For instance, the number of markers used for eye color prediction could be reduced to four while still maintaining reasonable predictive accuracy within our population.

CONCLUSIONS

Overall, our results suggest that it may be possible to reduce the number of SNPs used in some cases without compromising accuracy. However, further validation in larger and more diverse populations is essential to draw firm conclusions and make broader generalizations.

Survival in Patients with Uveal Melanoma Is Linked to Genetic Variation at HERC2 Single Nucleotide Polymorphism rs12913832

PURPOSE

Uveal melanoma (UM) is a rare disease, with the highest incidence in people with fair skin and light eyes. Eye color is largely genetically determined and is defined by a set of single nucleotide polymorphisms (SNPs). We set out to determine whether we could identify a SNP related to prognosis.

DESIGN

We sequenced DNA from peripheral blood mononuclear cells of 392 patients with UM and obtained the genotype of 6 common eye color-related SNPs. Clinical and histopathologic tumor characteristics, tumor chromosome status, and patient survival were compared among patients with different genotypes.

PARTICIPANTS

Three hundred ninety-two patients who underwent enucleation for UM at the Leiden University Medical Center, Leiden, The Netherlands.

METHODS

We isolated DNA from peripheral blood leukocytes of 392 patients with UM and performed sequencing, using 6 eye color SNPs from the HIrisPlex-S assay (Erasmus MC, Walsh lab). The genotypes extracted from the sequencing data were uploaded onto the HIrisPlexwebtool (https://hirisplex.erasmusmc.nl/) for eye color prediction. We tested the association of eye color SNPs with tumor characteristics and chromosome aberrations using Pearson's chi-square test and the Mann-Whitney U test and evaluated survival with Kaplan-Meier curves with the log-rank test and Cox regression.

MAIN OUTCOME MEASURES

Uveal melanoma-related survival.

RESULTS

Of 392 patients with analyzable genotype data, 307 patients (78%) were assigned blue eyes, 74 patients (19%) were assigned brown eyes, and 11 patients (3%) could not be assigned to either blue or brown. Patients with a genetically blue eye color showed worse survival (P = 0.04). This was related to 1 genotype: patients with the G/G genotype of rs12913832 (HERC2), which codes for blue eye color showed a worse prognosis (P = 0.017) and more often had high-risk tumors (monosomy of chromosome 3; P = 0.04) than in patients with an A/G or A/A genotype.

CONCLUSIONS

The G/G genotype of rs12913832 (HERC2), which is related to blue eye color, not only is a genetic factor related to the risk of UM develop, but also is linked to a worse prognosis because of an association with a higher risk of a high-risk UM developing (carrying monosomy of chromosome 3).

FINANCIAL DISCLOSURE(S)

The author(s) have no proprietary or commercial interest in any materials discussed in this article.

The Chromatin Organization Close to SNP rs12913832, Involved in Eye Color Variation, Is Evolutionary Conserved in Vertebrates

The most significant genetic influence on eye color pigmentation is attributed to the intronic SNP rs12913832 in the HERC2 gene, which interacts with the promoter region of the contiguous OCA2 gene. This interaction, through the formation of a chromatin loop, modulates the transcriptional activity of OCA2, directly affecting eye color pigmentation. Recent advancements in technology have elucidated the precise spatial organization of the genome within the cell nucleus, with chromatin architecture playing a pivotal role in regulating various genome functions. In this study, we investigated the organization of the chromatin close to the HERC2/OCA2 locus in human lymphocyte nuclei using fluorescence in situ hybridization (FISH) and high-throughput chromosome conformation capture (Hi-C) data. The 3 Mb of genomic DNA that belonged to the chromosomal region 15q12-q13.1 revealed the presence of three contiguous chromatin loops, which exhibited a different level of compaction depending on the presence of the A or G allele in the SNP rs12913832. Moreover, the analysis of the genomic organization of the genes has demonstrated that this chromosomal region is evolutionarily highly conserved, as evidenced by the analysis of syntenic regions in species from other Vertebrate classes. Thus, the role of rs12913832 variant is relevant not only in determining the transcriptional activation of the OCA2 gene but also in the chromatin compaction of a larger region, underscoring the critical role of chromatin organization in the proper regulation of the involved genes. It is crucial to consider the broader implications of this finding, especially regarding the potential regulatory role of similar polymorphisms located within intronic regions, which do not influence the same gene by modulating the splicing process, but they regulate the expression of adjacent genes. Therefore, caution should be exercised when utilizing whole-exome sequencing for diagnostic purposes, as intron sequences may provide valuable gene regulation information on the region where they reside. Thus, future research efforts should also be directed towards gaining a deeper understanding of the precise mechanisms underlying the role and mode of action of intronic SNPs in chromatin loop organization and transcriptional regulation.

Machine learning applications in forensic DNA profiling: A critical review

Machine learning (ML) is a range of powerful computational algorithms capable of generating predictive models via intelligent autonomous analysis of relatively large and often unstructured data. ML has become an integral part of our daily lives with a plethora of applications, including web, business, automotive industry, clinical diagnostics, scientific research, and more recently, forensic science. In the field of forensic DNA, the manual analysis of complex data can be challenging, time-consuming, and error-prone. The integration of novel ML-based methods may aid in streamlining this process while maintaining the high accuracy and reproducibility required for forensic tools. Due to the relative novelty of such applications, the forensic community is largely unaware of ML capabilities and limitations. Furthermore, computer science and ML professionals are often unfamiliar with the forensic science field and its specific requirements. This manuscript offers a brief introduction to the capabilities of machine learning methods and their applications in the context of forensic DNA analysis and offers a critical review of the current literature in this rapidly developing field.

Uveal melanoma: Current evidence on prognosis, treatment and potential developments

Uveal Melanoma (UM) is a rare disease, yet it is the most common primary intraocular malignancy in adult patients. Despite continuous advancements and research, the risk of metastasis remains high. It is possible to stratify patients according to their risk of metastases using a variety of known risk factors. Even though there is no gold standard for the prognostication of patients with uveal melanoma, it is becoming increasingly clear that combining histo-pathological, patient-related and molecular prognostic markers allows a more accurate prediction of the metastatic risk than by using one parameter. Primary UM in the eye are treated very effectively with eye-sparing radiation-based techniques or enucleation. However, it is not yet possible to prevent or treat metastases with the current therapeutic options. Nonetheless, the efforts to find new therapeutic targets continue and progress is being made, especially in the field of targeted therapy, as exemplified by the anti-gp100 bispecific molecule Tebentafusp. This review delves into the history of uveal melanoma, its incidence, presentation and diagnosis, the known prognostic factors and the treatment options, both for the primary tumour and for metastases. We show that different populations may have different risks for developing UM, and that each country should evaluate their own patients.

More than a decade of genetic research on the Denisovans

Denisovans, a group of now extinct humans who lived in Eastern Eurasia in the Middle and Late Pleistocene, were first identified from DNA sequences just over a decade ago. Only ten fragmentary remains from two sites have been attributed to Denisovans based entirely on molecular information. Nevertheless, there has been great interest in using genetic data to understand Denisovans and their place in human history. From the reconstruction of a single high-quality genome, it has been possible to infer their population history, including events of admixture with other human groups. Additionally, the identification of Denisovan DNA in the genomes of present-day individuals has provided insights into the timing and routes of dispersal of ancient modern humans into Asia and Oceania, as well as the contributions of archaic DNA to the physiology of present-day people. In this Review, we synthesize more than a decade of research on Denisovans, reconcile controversies and summarize insights into their population history and phenotype. We also highlight how our growing knowledge about Denisovans has provided insights into our own evolutionary history.

Genetic landscape of forensic DNA phenotyping markers among Mediterranean populations

Forensic DNA Phenotyping can reveal the appearance of an unknown individual by predicting the External Visible Characteristics (EVC) from DNA obtained at the crime scene. Our aim is to characterize the genetic landscape of Human identification markers responsible for EVC among Mediterranean populations compared to other worldwide groups. We conducted an exhaustive search for genes involved in EVC variation. Then, variants located on these genes were extracted from public genotypic data of Mediterranean, American, African and East Asiatic populations. The genetic landscape of these Human identification markers, their allelic distribution and admixture analyses, were determined using plink, R and ADMIXTURE softwares. Our results showed that the Mediterranean populations appear close to the Mexican populations and distinguished from sub Saharan African populations living in the USA and from East Asiatic populations. We highlighted a total of 103454 common variants shared between the studied populations and among them, 25 common variants associated with EVC. Interestingly, genotype frequencies results showed that the rs17646946, rs13016869, rs977588, rs1805008 and rs2240751 variants located respectively in the TCHH, PRKCE, OCA2, MC1R and MFSD12 genes are significantly different between the Mediterranean and Asiatic populations. The genotype frequencies of the variants rs977589 and rs7179994 located in the OCA2 gene, and of rs12913832 and rs2240751 located respectively in HERC2 and MFSD12 genes are significantly different between the Mediterranean and American populations. Our work generates a large number of EVC variants that could be a valuable resource for future studies in the forensic field.

The face as folded object: Race and the problems with 'progress' in forensic DNA phenotyping

Forensic DNA phenotyping (FDP) encompasses a set of technologies aimed at predicting phenotypic characteristics from genotypes. Advocates of FDP present it as the future of forensics, with an ultimate goal of producing complete, individualised facial composites based on DNA. With a focus on individuals and promised advances in technology comes the assumption that modern methods are steadily moving away from racial science. Yet in the quantification of physical differences, FDP builds upon some nineteenth- and twentieth-century scientific practices that measured and categorised human variation in terms of race. In this article I complicate the linear temporal approach to scientific progress by building on the notion of the folded object. Drawing on ethnographic fieldwork conducted in various genetic laboratories, I show how nineteenth- and early twentieth-century anthropological measuring and data-collection practices and statistical averaging techniques are folded into the ordering of measurements of skin color data taken with a spectrophotometer, the analysis of facial shape based on computational landmarks and the collection of iris photographs. Attending to the historicity of FDP facial renderings, I bring into focus how race comes about as a consequence of temporal folds.

Forensic DNA Phenotyping: Genes and Genetic Variants for Eye Color Prediction

In recent decades, the use of genetic polymorphisms related to specific phenotypes, such as eye color, has greatly contributed to the development of the research field called forensic DNA phenotyping (FDP), enabling the investigators of crime cases to reduce the number of suspects, making their work faster and more precise. Eye color is a polygenic phenotype, and many genetic variants have been highlighted, with the major contributor being the HERC2-OCA2 locus, where many single nucleotide variations (SNPs) were identified. Interestingly, the HERC2-OCA2 locus, containing the intronic SNP rs12913832, the major eye color determinant, shows a high level of evolutionary conservation across many species of vertebrates. Currently, there are some genetic panels to predict eye color by genomic DNA analysis, even if the exact role of the SNP variants in the formation of eye color is still poorly understood, with a low level of predictivity in the so-called intermediate eye color. Many variants in OCA2, HERC2, and other genes lie in introns or correspond to synonymous variants, highlighting greater complexity in the mechanism of action of such genes than a simple missense variation. Here, we show the main genes involved in oculocutaneous pigmentation and their structural and functional features, as well as which genetic variants show the highest level of eye color predictivity in currently used FDP assays. Despite the great recent advances and impact of FDP in criminal cases, it is necessary to enhance scientific research to better understand the mechanism of action behind each genetic variant involved in eye color, with the goal of obtaining higher levels of prediction.

Phenotypic Classification of Eye Colour and Developmental Validation of the Irisplex System on Population Living in Malakand Division, Pakistan

The core objective of forensic DNA typing is developing DNA profiles from biological evidence for personal identification. The present study was designed to check the validation of the IrisPlex system and the Prevalence of eye colour in the Pakhtoon population residing within the Malakand Division.

METHODS

Eye colour digital photographs and buccal swab samples of 893 individuals of different age groups were collected. Multiplexed SNaPshot single base extension chemistry was used, and the genotypic results were analysed. Snapshot data were used for eye colour prediction through the IrisPlex and FROG-kb tool.

RESULTS

The results of the present study found brown eye colour to be the most prevalent eye colour in comparison to intermediate and blue coloured. Overall, individuals with brown-coloured eyes possess CT (46.84%) and TT (53.16%) genotypes. Blue eye-coloured individuals are solely of the CC genotype, while individuals of intermediate eye colour carry CT (45.15%) and CC (53.85%) genotypes in rs12913832 SNP in the HERC2 gene. It was also revealed that brown-coloured eyes individuals were dominant among all age groups followed by intermediate and blue. Statistical analysis between particular variables and eye colour showed a significant p-value (<0.05) for rs16891982 SNP in SLC45A2 gene, rs12913832 SNP in HERC2 gene, rs1393350 SNP in SLC45A2, districts and gender. The rest of the SNPs were non-significant with eye colour, respectively. The rs12896399 SNP and SNP rs1800407 were found significant with rs16891982 SNP. The result also demonstrated that the study group differs from the world population based on eye colour. The two eye colour prediction results were compared, and it was discovered that IrisPlex and FROG-Kb had similar higher prediction ratios for Brown and Blue eye colour.

CONCLUSIONS

The results of the current study revealed brown eye colour to be the most prevalent amongst members of the local population of Pakhtoon ethnicity in the Malakand Division of northern Pakistan. A set of contemporary human DNA samples with known phenotypes are used in this research to evaluate the custom panel's prediction accuracy. With the aid of this forensic test, DNA typing can be supplemented with details about the appearance of the person from whom the sample was taken in cases involving missing persons, ancient human remains, and trace samples. This study may be helpful for future population genetics and forensics studies.

Recent advances in Forensic DNA Phenotyping of appearance, ancestry and age

Forensic DNA Phenotyping (FDP) comprises the prediction of a person's externally visible characteristics regarding appearance, biogeographic ancestry and age from DNA of crime scene samples, to provide investigative leads to help find unknown perpetrators that cannot be identified with forensic STR-profiling. In recent years, FDP has advanced considerably in all of its three components, which we summarize in this review article. Appearance prediction from DNA has broadened beyond eye, hair and skin color to additionally comprise other traits such as eyebrow color, freckles, hair structure, hair loss in men, and tall stature. Biogeographic ancestry inference from DNA has progressed from continental ancestry to sub-continental ancestry detection and the resolving of co-ancestry patterns in genetically admixed individuals. Age estimation from DNA has widened beyond blood to more somatic tissues such as saliva and bones as well as new markers and tools for semen. Technological progress has allowed forensically suitable DNA technology with largely increased multiplex capacity for the simultaneous analysis of hundreds of DNA predictors with targeted massively parallel sequencing (MPS). Forensically validated MPS-based FDP tools for predicting from crime scene DNA i) several appearance traits, ii) multi-regional ancestry, iii) several appearance traits together with multi-regional ancestry, and iv) age from different tissue types, are already available. Despite recent advances that will likely increase the impact of FDP in criminal casework in the near future, moving reliable appearance, ancestry and age prediction from crime scene DNA to the level of detail and accuracy police investigators may desire, requires further intensified scientific research together with technical developments and forensic validations as well as the necessary funding.

Human genetic history on the Tibetan Plateau in the past 5100 years

Using genome-wide data of 89 ancient individuals dated to 5100 to 100 years before the present (B.P.) from 29 sites across the Tibetan Plateau, we found plateau-specific ancestry across plateau populations, with substantial genetic structure indicating high differentiation before 2500 B.P. Northeastern plateau populations rapidly showed admixture associated with millet farmers by 4700 B.P. in the Gonghe Basin. High genetic similarity on the southern and southwestern plateau showed population expansion along the Yarlung Tsangpo River since 3400 years ago. Central and southeastern plateau populations revealed extensive genetic admixture within the plateau historically, with substantial ancestry related to that found in southern and southwestern plateau populations. Over the past ~700 years, substantial gene flow from lowland East Asia further shaped the genetic landscape of present-day plateau populations. The high-altitude adaptive EPAS1 allele was found in plateau populations as early as in a 5100-year-old individual and showed a sharp increase over the past 2800 years.

Association between Variants in the OCA2-HERC2 Region and Blue Eye Colour in HERC2 rs12913832 AA and AG Individuals

The OCA2-HERC2 region is strongly associated with human pigmentation, especially eye colour. The HERC2 SNP rs12913832 is currently the best-known predictor for blue and brown eye colour. However, in a previous study we found that 43 of 166 Norwegians with the brown eye colour genotype rs12913832:AA or AG, did not have the expected brown eye colour. In this study, we carried out massively parallel sequencing of a ~500 kbp HERC2-OCA2 region in 94 rs12913832:AA and AG Norwegians (43 blue-eyed and 51 brown-eyed) to search for novel blue eye colour variants. The new candidate variants were subsequently typed in a Norwegian biobank population (total n = 519) for population specific association analysis. We identified five new variants, rs74409036:A, rs78544415:T, rs72714116:T, rs191109490:C and rs551217952:C, to be the most promising candidates for explaining blue eye colour in individuals with the rs12913832:AA and AG genotype. Additionally, we confirmed the association of the missense variants rs74653330:T and rs121918166:T with blue eye colour, and observed lighter skin colour in rs74653330:T individuals. In total, 37 (86%) of the 43 blue-eyed rs12913832:AA and AG Norwegians could potentially be explained by these seven variants, and we suggest including them in future prediction models.

Increased histological tumour pigmentation in Uveal Melanoma is related to eye colour and loss of chromosome 3/BAP1

Purpose

Heavy pigmentation is known to be a prognostic risk factor in uveal melanoma (UM). We analyzed whether genetic tumor parameters were associated with tumor pigmentation and whether pigmentation should be included in prognostic tests.

Design

Retrospective comparison of clinical, histopathological, and genetic features and survival in UM with different pigmentation.

Participants

A total of 1058 patients with UM from a White European population with diverse eye colors enucleated between 1972 and 2021.

Methods

Cox regression and log-rank tests were used for survival analysis; the chi-square test and Mann-Whitney U test were used for correlation analysis.

Main Outcome Measures

Uveal melanoma-related survival based on tumor pigmentation and chromosome status, correlation of tumor pigmentation with prognostic factors.

Results

The 5-year UM-related mortality was 8% in patients with nonpigmented tumors (n = 54), 25% with lightly pigmented tumors (n = 489), 41% with moderately pigmented tumors (n = 333), and 33% with dark tumors (n = 178) (P < 0.001). The percentage of tumors with monosomy 3 (M3) or 8q gain increased with increasing pigmentation (31%, 46%, 62%, and 70% having M3 [P < 0.001], and 19%, 43%, 61%, and 63% having 8q gain [P < 0.001] in the 4 increasing pigment groups, respectively). BRCA-associated protein 1 (BAP1) loss (known for 204 cases) was associated with increased tumor pigmentation (P = 0.001). Cox regression analysis on survival showed that when chromosome status and pigmentation were both included, pigmentation was not an independent prognostic indicator. Preferentially expressed antigen in melanoma (PRAME) expression was a significant prognostic marker in light tumors (P = 0.02) but not in dark tumors (P = 0.85).

Conclusions

Patients with moderately and heavily pigmented tumors showed a significantly higher UM-related mortality than patients with unpigmented and light tumors (P < 0.001), supporting prior reports on the relation between increased tumor pigmentation and a worse prognosis. Although we previously showed that a dark eye color was associated with tumor pigmentation, we now show that the tumor's genetic status (chromosome 3 and 8q/BAP1 status) is also related to tumor pigmentation. When pigmentation and chromosome 3 status are both included in a Cox regression analysis, pigmentation is not an independent prognostic factor. However, evidence from this and previous studies shows that chromosome changes and PRAME expression have a stronger association with survival when they occur in light tumors than in dark ones. Financial Disclosure(s): Proprietary or commercial disclosure may be found after the references.

Advancement in Human Face Prediction Using DNA

The rapid improvements in identifying the genetic factors contributing to facial morphology have enabled the early identification of craniofacial syndromes. Similarly, this technology can be vital in forensic cases involving human identification from biological traces or human remains, especially when reference samples are not available in the deoxyribose nucleic acid (DNA) database. This review summarizes the currently used methods for predicting human phenotypes such as age, ancestry, pigmentation, and facial features based on genetic variations. To identify the facial features affected by DNA, various two-dimensional (2D)- and three-dimensional (3D)-scanning techniques and analysis tools are reviewed. A comparison between the scanning technologies is also presented in this review. Face-landmarking techniques and face-phenotyping algorithms are discussed in chronological order. Then, the latest approaches in genetic to 3D face shape analysis are emphasized. A systematic review of the current markers that passed the threshold of a genome-wide association (GWAS) of single nucleotide polymorphism (SNP)-face traits from the GWAS Catalog is also provided using the preferred reporting items for systematic reviews and meta-analyses (PRISMA), approach. Finally, the current challenges in forensic DNA phenotyping are analyzed and discussed.

Predicting Eye and Hair Color in a Turkish Population Using the HIrisPlex System

Forensic DNA Phenotyping (FDP) can reveal the appearance of an unknown individual by predicting the ancestry, phenotype (i.e., hair, eye, skin color), and age from DNA obtained at the crime scene. The HIrisPlex system has been developed to simultaneously predict eye and hair color. However, the prediction accuracy of the system needs to be assessed for the tested population before implementing FDP in casework. In this study, we evaluated the performance of the HIrisPlex system on 149 individuals from the Turkish population. We applied the single-based extension (SNaPshot chemistry) method and used the HIrisPlex online tool to test the prediction of the eye and hair colors. The accuracy of the HIrisPlex system was assessed through the calculation of the area under the receiver characteristic operating curves (AUC), sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV). The results showed that the proposed method successfully predicted the eye and hair color, especially for blue (100%) and brown (95.60%) eye and black (95.23) and brown (98.94) hair colors. As observed in previous studies, the system failed to predict intermediate eye color, representing 25% in our cohort. The majority of incorrect predictions were observed for blond hair color (40.7%). Previous HIrisPlex studies have also noted difficulties with these phenotypes. Our study shows that the HIrisPlex system can be applied to forensic casework in Turkey with careful interpretation of the data, particularly intermediate eye color and blond hair color.

Eye and Hair Color Prediction of Ancient and Second World War Skeletal Remains Using a Forensic PCR-MPS Approach

To test the usefulness of the forensic PCR-MPS approach to eye and hair color prediction for aged skeletons, a customized version of the PCR-MPS HIrisPlex panel was used on two sets of samples. The first set contained 11 skeletons dated from the 3rd to the 18th centuries AD, and for each of them at least four bone types were analyzed (for a total of 47 samples). In the second set, 24 skeletons from the Second World War were analyzed, and only petrous bones from the skulls were tested. Good-quality libraries were achieved in 83.3% of the cases for the ancient skeletons and in all Second World War petrous bones, with 94.7% and 100% of the markers, respectively, suitable for SNP typing. Consensus typing was achieved for about 91.7% of the markers in 10 out of 11 ancient skeletons, and the HIrisPlex-S webtool was then used to generate phenotypic predictions. Full predictions were achieved for 3 (27.3%) ancient skeletons and 12 (50%) Second World War petrous bones. In the remaining cases, different levels of AUC (area under the receiver operating curve) loss were computed because of no available data (NA) for 8.3% of markers in ancient skeletons and 4.2% of markers in Second World War petrous bones. Although the PCR-based approach has been replaced with new techniques in ancient DNA studies, the results show that customized forensic technologies can be successfully applied to aged bone remains, highlighting the role of the template in the success of PCR-MPS analysis. However, because several typical errors of ancient DNA sequencing were scored, replicate tests and accurate evaluation by an expert remain indispensable tools.

Applicability of the IrisPlex system for eye color prediction in an admixed population from Argentina

Eye color prediction based on an individual's genetic information is of interest in the field of forensic genetics. In recent years, researchers have studied different genes and markers associated with this externally visible characteristic and have developed methods for its prediction. The IrisPlex represents a validated tool for homogeneous populations, though its applicability in populations of mixed ancestry is limited, mainly regarding the prediction of intermediate eye colors. With the aim of validating the applicability of this system in an admixed population from Argentina (n = 302), we analyzed the six single nucleotide variants used in that multiplex for eye color and four additional SNPs, and evaluated its prediction ability. We also performed a genotype-phenotype association analysis. This system proved to be useful when dealing with the extreme ends of the eye color spectrum (blue and brown) but presented difficulties in determining the intermediate phenotypes (green), which were found in a large proportion of our population. We concluded that these genetic tools should be used with caution in admixed populations and that more studies are required in order to improve the prediction of intermediate phenotypes.

MITF in Normal Melanocytes, Cutaneous and Uveal Melanoma: A Delicate Balance

Microphthalmia-associated transcription factor (MITF) is an important regulator of melanogenesis and melanocyte development. Although it has been studied extensively in cutaneous melanoma, the role of MITF in uveal melanoma (UM) has not been explored in much detail. We review the literature about the role of MITF in normal melanocytes, in cutaneous melanoma, and in UM. In normal melanocytes, MITF regulates melanocyte development, melanin synthesis, and melanocyte survival. The expression profile and the behaviour of MITF-expressing cells suggest that MITF promotes local proliferation and inhibits invasion, inflammation, and epithelial-to-mesenchymal (EMT) transition. Loss of MITF expression leads to increased invasion and inflammation and is more prevalent in malignant cells. Cutaneous melanoma cells switch between MITF-high and MITF-low states in different phases of tumour development. In UM, MITF loss is associated with loss of BAP1 protein expression, which is a marker of poor prognosis. These data indicate a dual role for MITF in benign and malignant melanocytic cells.

Predicting Physical Appearance from DNA Data-Towards Genomic Solutions

The idea of forensic DNA intelligence is to extract from genomic data any information that can help guide the investigation. The clues to the externally visible phenotype are of particular practical importance. The high heritability of the physical phenotype suggests that genetic data can be easily predicted, but this has only become possible with less polygenic traits. The forensic community has developed DNA-based predictive tools by employing a limited number of the most important markers analysed with targeted massive parallel sequencing. The complexity of the genetics of many other appearance phenotypes requires big data coupled with sophisticated machine learning methods to develop accurate genomic predictors. A significant challenge in developing universal genomic predictive methods will be the collection of sufficiently large data sets. These should be created using whole-genome sequencing technology to enable the identification of rare DNA variants implicated in phenotype determination. It is worth noting that the correctness of the forensic sketch generated from the DNA data depends on the inclusion of an age factor. This, however, can be predicted by analysing epigenetic data. An important limitation preventing whole-genome approaches from being commonly used in forensics is the slow progress in the development and implementation of high-throughput, low DNA input sequencing technologies. The example of palaeoanthropology suggests that such methods may possibly be developed in forensics.

Further insight into the global variability of the OCA2-HERC2 locus for human pigmentation from multiallelic markers

The OCA2-HERC2 locus is responsible for the greatest proportion of eye color variation in humans. Numerous studies extensively described both functional SNPs and associated patterns of variation over this region. The goal of our study is to examine how these haplotype structures and allelic associations vary when highly variable markers such as microsatellites are used. Eleven microsatellites spanning 357 Kb of OCA2-HERC2 genes are analyzed in 3029 individuals from worldwide populations. We found that several markers display large differences in allele frequency (10% to 35% difference) among Europeans, East Asians and Africans. In Europe, the alleles showing increased frequency can also discriminate individuals with (IrisPlex) predicted blue and brown eyes. Distinct haplotypes are identified around the variants C and T of the functional SNP rs12913832 (associated to blue eyes), with linkage disequilibrium r² values significant up to 237 Kb. The haplotype carrying the allele rs12913832 C has high frequency (76%) in blue eye predicted individuals (30% in brown eye predicted individuals), while the haplotype associated to the allele rs12913832 T is restricted to brown eye predicted individuals. Finally, homozygosity values reach levels of 91% near rs12913832. Odds ratios show values of 4.2, 7.4 and 10.4 for four markers around rs12913832 and 7.1 for their core haplotype. Hence, this study provides an example on the informativeness of multiallelic markers that, despite their current limited potential contribution to forensic eye color prediction, supports the use of microsatellites for identifying causing variants showing similar genetic features and history.

Chromosome 3 and 8q aberrations in Uveal Melanoma show greater impact on survival in patients with light iris versus dark iris color

PURPOSE

Individuals with gray, blue, or green iris have a higher chance of developing uveal melanoma (UM) than those with brown eyes. We wondered whether iris pigmentation might not only be related to predisposition to UM, but also to its behavior and therefore compared clinical, histopathologic, and genetic characteristics of UM between eyes with different iris colors.

DESIGN

We determined iris color in a large cohort of patients who had undergone an enucleation for UM. Clinical and histopathological tumor characteristics, chromosome status, and survival were compared between three groups, based on iris color.

PARTICIPANTS

412 patients with choroidal/ciliary body UM, who had undergone primary enucleation at the Leiden University Medical Center (LUMC), Leiden, The Netherlands, between 1993 and 2019, divided into three groups, based on iris color (gray/blue, green/hazel, and brown). Validation cohort: 934 choroidal/ciliary body UM patients treated at Wills Eye Hospital (WEH), Philadelphia, United States.

METHODS

Comparison of clinical, histopathologic, and genetic characteristics of UM in patients with different iris colors.

MAIN OUTCOME MEASURES

Melanoma-related survival in UM patients, divided over three iris color groups, in relation to the tumor's chromosome 3 and 8q status.

RESULTS

Moderate and heavy tumor pigmentation was especially seen in eyes with brown iris (p < 0.001). Survival did not differ between patients with different iris colors (p = 0.28). However, in patients with a light iris, copy number changes in chromosome 3 and 8q had a greater influence on survival than in patients with a dark iris. Similarly, chromosome 3 and chromosome 8q status affected survival more among patients with lightly-pigmented tumors than in patients with heavily-pigmented tumors. The WEH cohort similarly showed a greater influence of chromosome aberrations in light-eyed individuals.

CONCLUSIONS

While iris color by itself did not relate to survival of UM patients, chromosome 3 and 8q aberrations had a much larger influence on survival in patients with light iris compared to those with brown iris. This suggests a synergistic effect of iris pigmentation and chromosome status in the regulation of oncogenic behavior of UM. Iris color should be taken into consideration when calculating the risk for developing metastases.

Predicting eye and hair colour in a Norwegian population using Verogen's ForenSeq™ DNA signature prep kit

Prediction of eye and hair colour from DNA can be an important investigative tool in forensic cases if conventional DNA profiling fails to match DNA from any known suspects or cannot obtain a hit in a DNA database. The HIrisPlex model for simultaneous eye and hair colour predictions was developed for forensic usage. To genotype a DNA sample, massively parallel sequencing (MPS) has brought new possibilities to the analysis of forensic DNA samples. As part of an in-house validation, this study presents the genotyping and predictive performance of the HIrisPlex SNPs in a Norwegian study population, using Verogen's ForenSeq™ DNA Signature Prep Kit on the MiSeq FGx system and the HIrisPlex webtool. DNA-profiles were successfully typed with DNA input down to 125 pg. In samples with DNA input < 125 pg, false homozygotes were observed with as many as 92 reads. Prediction accuracies in terms of AUC were high for red (0.97) and black (0.93) hair colours, as well as blue (0.85) and brown (0.94) eye colours. The AUCs for blond (0.72) and brown (0.70) hair colour were considerably lower. None of the individuals was predicted to have intermediate eye colour. Therefore, the error rates of the overall eye colour predictions were 37% with no predictive probability threshold (pmax) and 26% with a probability threshold of 0.7. We also observed that more than half of the incorrect predictions were for individuals carrying the rs12913832 GG genotype. For hair colour, 65% of the individuals were correctly predicted when using the highest probability category approach. The main error was observed for individuals with brown hair colour that were predicted to have blond hair. Utilising the prediction guide approach increased the correct predictions to 75%. Assessment of phenotype-genotype associations of eye colours using a quantitative eye colour score (PIE-score), revealed that rs12913832 AA individuals of Norwegian descent had statistically significantly higher PIE-score (less brown eye colour) than individuals of non-northern European descent. To our knowledge, this has not been reported in other studies. Our study suggests that careful assessment of the target population prior to the implementation of forensic DNA phenotyping to case work is beneficial.

GenNet framework: interpretable deep learning for predicting phenotypes from genetic data

Applying deep learning in population genomics is challenging because of computational issues and lack of interpretable models. Here, we propose GenNet, a novel open-source deep learning framework for predicting phenotypes from genetic variants. In this framework, interpretable and memory-efficient neural network architectures are constructed by embedding biologically knowledge from public databases, resulting in neural networks that contain only biologically plausible connections. We applied the framework to seventeen phenotypes and found well-replicated genes such as HERC2 and OCA2 for hair and eye color, and novel genes such as ZNF773 and PCNT for schizophrenia. Additionally, the framework identified ubiquitin mediated proteolysis, endocrine system and viral infectious diseases as most predictive biological pathways for schizophrenia. GenNet is a freely available, end-to-end deep learning framework that allows researchers to develop and use interpretable neural networks to obtain novel insights into the genetic architecture of complex traits and diseases.

van Hilten and colleagues present GenNet, a deep-learning framework for predicting phenotype from genetic data. This framework generates interpretable neural networks that provide insight into the genetic basis of complex traits and diseases.

Pushing the Boundaries: Forensic DNA Phenotyping Challenged by Single-Cell Sequencing

Single-cell sequencing is a fast developing and very promising field; however, it is not commonly used in forensics. The main motivation behind introducing this technology into forensics is to improve mixture deconvolution, especially when a trace consists of the same cell type. Successful studies demonstrate the ability to analyze a mixture by separating single cells and obtaining CE-based STR profiles. This indicates a potential use of the method in other forensic investigations, like forensic DNA phenotyping, in which using mixed traces is not fully recommended. For this study, we collected single-source autopsy blood from which the white cells were first stained and later separated with the DEPArray™ N×T System. Groups of 20, 10, and 5 cells, as well as 20 single cells, were collected and submitted for DNA extraction. Libraries were prepared using the Ion AmpliSeq™ PhenoTrivium Panel, which includes both phenotype (HIrisPlex-S: eye, hair, and skin color) and ancestry-associated SNP-markers. Prior to sequencing, half of the single-cell-based libraries were additionally amplified and purified in order to improve the library concentrations. Ancestry and phenotype analysis resulted in nearly full consensus profiles resulting in correct predictions not only for the cells groups but also for the ten re-amplified single-cell libraries. Our results suggest that sequencing of single cells can be a promising tool used to deconvolute mixed traces submitted for forensic DNA phenotyping.

Reliability of phenotype estimation and extended classification of ancestry using decedent samples

The Illumina® MiSeq FGx™, in conjunction with the ForenSeq™ DNA Signature Prep kit, produces genotypes of the CODIS-required short tandem repeats and provides phenotype and biogeographical ancestry estimations via phenotype-informative and ancestry-informative markers, respectively. Although both markers have been validated for use in forensic biology, there is little data to determine the practical utility of these estimations to assist in identifying missing persons using decedent casework samples. The accuracy and utility of phenotypic and ancestral estimations were investigated for 300 samples received by the Los Angeles County Department of Medical Examiner-Coroner. piSNP genotypes were translated into hair and eye colors using the Forenseq™ Universal Analysis Software (UAS) on the MiSeq FGx™ and the HIrisPlex System, and statistical accuracy was evaluated in context with the reported decedent characteristics. Similarly, estimates of each decedent's biogeographical ancestry were compared to assess the efficacy of these markers to predict ancestry correctly. The average UAS and the HIrisPlex system prediction accuracy for brown and blue eyes were 95.3% and 96.2%, respectively. Intermediate eye color could not be predicted with high accuracy using either system. Other than the black hair phenotype reporting an accuracy that exceeded 90% using either system, hair color was also too variable to be predicted with high accuracy. The FROG-kb database distinguishes decedents adequately beyond the Asian, African, European, and Admixed American global ancestries provided by the MiSeq FGx™ UAS PCA plots. FROG-kb correctly identified Middle Eastern, Pacific Islander, Latin American, or Jewish ancestries with accuracies of 70.0%, 81.8%, 73.8%, and 86.7%, respectively.

Searching for improvements in predicting human eye colour from DNA

Increasing understanding of human genome variability allows for better use of the predictive potential of DNA. An obvious direct application is the prediction of the physical phenotypes. Significant success has been achieved, especially in predicting pigmentation characteristics, but the inference of some phenotypes is still challenging. In search of further improvements in predicting human eye colour, we conducted whole-exome (enriched in regulome) sequencing of 150 Polish samples to discover new markers. For this, we adopted quantitative characterization of eye colour phenotypes using high-resolution photographic images of the iris in combination with DIAT software analysis. An independent set of 849 samples was used for subsequent predictive modelling. Newly identified candidates and 114 additional literature-based selected SNPs, previously associated with pigmentation, and advanced machine learning algorithms were used. Whole-exome sequencing analysis found 27 previously unreported candidate SNP markers for eye colour. The highest overall prediction accuracies were achieved with LASSO-regularized and BIC-based selected regression models. A new candidate variant, rs2253104, located in the ARFIP2 gene and identified with the HyperLasso method, revealed predictive potential and was included in the best-performing regression models. Advanced machine learning approaches showed a significant increase in sensitivity of intermediate eye colour prediction (up to 39%) compared to 0% obtained for the original IrisPlex model. We identified a new potential predictor of eye colour and evaluated several widely used advanced machine learning algorithms in predictive analysis of this trait. Our results provide useful hints for developing future predictive models for eye colour in forensic and anthropological studies.

Prediction of Eye Colour in Scandinavians Using the EyeColour 11 (EC11) SNP Set

Description of a perpetrator's eye colour can be an important investigative lead in a forensic case with no apparent suspects. Herein, we present 11 SNPs (Eye Colour 11-EC11) that are important for eye colour prediction and eye colour prediction models for a two-category reporting system (blue and brown) and a three-category system (blue, intermediate, and brown). The EC11 SNPs were carefully selected from 44 pigmentary variants in seven genes previously found to be associated with eye colours in 757 Europeans (Danes, Swedes, and Italians). Mathematical models using three different reporting systems: a quantitative system (PIE-score), a two-category system (blue and brown), and a three-category system (blue, intermediate, brown) were used to rank the variants. SNPs with a sufficient mean variable importance (above 0.3%) were selected for EC11. Eye colour prediction models using the EC11 SNPs were developed using leave-one-out cross-validation (LOOCV) in an independent data set of 523 Norwegian individuals. Performance of the EC11 models for the two- and three-category system was compared with models based on the IrisPlex SNPs and the most important eye colour locus, rs12913832. We also compared model performances with the IrisPlex online tool (IrisPlex Web). The EC11 eye colour prediction models performed slightly better than the IrisPlex and rs12913832 models in all reporting systems and better than the IrisPlex Web in the three-category system. Three important points to consider prior to the implementation of eye colour prediction in a forensic genetic setting are discussed: (1) the reference population, (2) the SNP set, and (3) the reporting strategy.

Validation of the Verogen ForenSeq™ DNA Signature Prep kit/Primer Mix B for phenotypic and biogeographical ancestry predictions using the Micro MiSeq® Flow Cells

In anticipation of offering phenotypic and biogeographical ancestry predictions to help resolve cases, the Verogen ForenSeq™ DNA Signature Prep kit/Primer Mix B was evaluated in the context of Micro MiSeq® Flow Cells. These flow cells were determined as the best format for a quick turnaround time response and cost effective approach compared to standard flow cells. The phenotype informative SNPs (piSNPs) and ancestry informative SNPs (aiSNPs) were thoroughly examined through sensitivity, reproducibility and repeatability, concordance, robustness (mock casework) and low level DNA mixture studies purposely selecting individuals with different phenotypes (hair and eye color) when possible and different biogeographical ancestry. SNP locus-specific interpretation thresholds were established for the Universal Analysis Software (UAS) based on surviving alleles and SNP predictor rank to minimize false homozygous genotypes and maximize the information that can be derived from an unknown sample. Dropin alleles' intensity determined an appropriate threshold to minimize false heterozygous SNP genotypes. The selection of inappropriate interpretation thresholds was shown to have major consequences on phenotypic predictions. A 3.2% and 4.8% minor DNA component contribution to a DNA mixture had no impact on ancestry predictions whereas a 9.1% contribution did. The multi-locus SNP genotypes generated using the ForenSeq™ DNA Signature Prep kit/Primer Mix B were shown to be reliable, reproducible, concordant and resulted in predictions that were also reliable, reproducible and concordant based on the limited number of donors (N = 19) used in this study.

Analysis of Skin Pigmentation and Genetic Ancestry in Three Subpopulations from Pakistan: Punjabi, Pashtun, and Baloch

Skin pigmentation is one of the most prominent and variable phenotypes in humans. We compared the alleles of 163 SNPs and indels from the Human Pigmentation (HuPi) AmpliSeq™ Custom panel, and biogeographic ancestry with the quantitative skin pigmentation levels on the upper arm, lower arm, and forehead of 299 Pakistani individuals from three subpopulations: Baloch, Pashtun, and Punjabi. The biogeographic ancestry of each individual was estimated using the Precision ID Ancestry Panel. All individuals were mainly of mixed South-Central Asian and European ancestry. However, the Baloch individuals also had an average proportion of Sub-Saharan African ancestry of approximately 10%, whereas it was <1% in the Punjabi and Pashtun individuals. The pairwise genetic distances between the Pashtun, Punjabi, and Baloch subpopulations based on the ancestry markers were statistically significantly different. Individuals from the Pashtun subpopulation had statistically significantly lower skin pigmentation than individuals from the Punjabi and Baloch subpopulations (p < 0.05). The proportions of European and Sub-Saharan African ancestry and five SNPs (rs1042602, rs10831496, rs1426654, rs16891982, and rs12913832) were statistically significantly associated with skin pigmentation at either the upper arm, lower arm or forehead in the Pakistani population after correction for multiple testing (p < 10⁻³). A model based on four of these SNPs (rs1426654, rs1042602, rs16891982, and rs12913832) explained 33% of the upper arm skin pigmentation. The four SNPs and the proportions of European and Sub-Saharan African ancestry explained 37% of the upper arm skin pigmentation. Our results indicate that the four likely causative SNPs, rs1426654, rs1042602, rs16891982, and rs12913832 located in SLC24A5, TYR, SLC45A2, and HERC2, respectively, are essential for skin color variation in the admixed Pakistani subpopulations.

Prediction of eye and hair pigmentation phenotypes using the HIrisPlex system in a Brazilian admixed population sample

Human pigmentation is a complex trait, probably involving more than 100 genes. Predicting phenotypes using SNPs present in those genes is important for forensic purpose. For this, the HIrisPlex tool was developed for eye and hair color prediction, with both models achieving high accuracy among Europeans. Its evaluation in admixed populations is important, since they present a higher frequency of intermediate phenotypes, and HIrisPlex has demonstrated limitations in such predictions; therefore, the performance of this tool may be impaired in such populations. Here, we evaluate the set of 24 markers from the HIrisPlex system in 328 individuals from Ribeirão Preto (SP) region, predicting eye and hair color and comparing the predictions with their real phenotypes. We used the HaloPlex Target Enrichment System and MiSeq Personal Sequencer platform for massively parallel sequencing. The prediction of eye and hair color was accomplished by the HIrisPlex online tool, using the default prediction settings. Ancestry was estimated using the SNPforID 34-plex to observe if and how an individual's ancestry background would affect predictions in this admixed sample. Our sample presented major European ancestry (70.5%), followed by African (21.1%) and Native American/East Asian (8.4%). HIrisPlex presented an overall sensitivity of 0.691 for hair color prediction, with sensitivities ranging from 0.547 to 0.782. The lowest sensitivity was observed for individuals with black hair, who present a reduced European contribution (48.4%). For eye color prediction, the overall sensitivity was 0.741, with sensitivities higher than 0.85 for blue and brown eyes, although it failed in predicting intermediate eye color. Such struggle in predicting this phenotype category is in accordance with what has been seen in previous studies involving HIrisPlex. Individuals with brown eye color are more admixed, with European ancestry decreasing to 62.6%; notwithstanding that, sensitivity for brown eyes was almost 100%. Overall sensitivity increases to 0.791 when a 0.7 threshold is set, though 12.5% of the individuals become undefined. When combining eye and hair prediction, hit rates between 51.3 and 68.9% were achieved. Despite the difficulties with intermediate phenotypes, we have shown that HIrisPlex results can be very helpful when interpreted with caution.

Eye-color and Type-2 diabetes phenotype prediction from genotype data using deep learning methods

Background

Genotype–phenotype predictions are of great importance in genetics. These predictions can help to find genetic mutations causing variations in human beings. There are many approaches for finding the association which can be broadly categorized into two classes, statistical techniques, and machine learning. Statistical techniques are good for finding the actual SNPs causing variation where Machine Learning techniques are good where we just want to classify the people into different categories. In this article, we examined the Eye-color and Type-2 diabetes phenotype. The proposed technique is a hybrid approach consisting of some parts from statistical techniques and remaining from Machine learning.

Results

The main dataset for Eye-color phenotype consists of 806 people. 404 people have Blue-Green eyes where 402 people have Brown eyes. After preprocessing we generated 8 different datasets, containing different numbers of SNPs, using the mutation difference and thresholding at individual SNP. We calculated three types of mutation at each SNP no mutation, partial mutation, and full mutation. After that data is transformed for machine learning algorithms. We used about 9 classifiers, RandomForest, Extreme Gradient boosting, ANN, LSTM, GRU, BILSTM, 1DCNN, ensembles of ANN, and ensembles of LSTM which gave the best accuracy of 0.91, 0.9286, 0.945, 0.94, 0.94, 0.92, 0.95, and 0.96% respectively. Stacked ensembles of LSTM outperformed other algorithms for 1560 SNPs with an overall accuracy of 0.96, AUC = 0.98 for brown eyes, and AUC = 0.97 for Blue-Green eyes. The main dataset for Type-2 diabetes consists of 107 people where 30 people are classified as cases and 74 people as controls. We used different linear threshold to find the optimal number of SNPs for classification. The final model gave an accuracy of 0.97%.

Conclusion

Genotype–phenotype predictions are very useful especially in forensic. These predictions can help to identify SNP variant association with traits and diseases. Given more datasets, machine learning model predictions can be increased. Moreover, the non-linearity in the Machine learning model and the combination of SNPs Mutations while training the model increases the prediction. We considered binary classification problems but the proposed approach can be extended to multi-class classification.

Prediction of eye, hair and skin colour in Latin Americans

Here we evaluate the accuracy of prediction for eye, hair and skin pigmentation in a dataset of > 6500 individuals from Mexico, Colombia, Peru, Chile and Brazil (including genome-wide SNP data and quantitative/categorical pigmentation phenotypes - the CANDELA dataset CAN). We evaluated accuracy in relation to different analytical methods and various phenotypic predictors. As expected from statistical principles, we observe that quantitative traits are more sensitive to changes in the prediction models than categorical traits. We find that Random Forest or Linear Regression are generally the best performing methods. We also compare the prediction accuracy of SNP sets defined in the CAN dataset (including 56, 101 and 120 SNPs for eye, hair and skin colour prediction, respectively) to the well-established HIrisPlex-S SNP set (including 6, 22 and 36 SNPs for eye, hair and skin colour prediction respectively). When training prediction models on the CAN data, we observe remarkably similar performances for HIrisPlex-S and the larger CAN SNP sets for the prediction of hair (categorical) and eye (both categorical and quantitative), while the CAN sets outperform HIrisPlex-S for quantitative, but not for categorical skin pigmentation prediction. The performance of HIrisPlex-S, when models are trained in a world-wide sample (although consisting of 80% Europeans, https://hirisplex.erasmusmc.nl), is lower relative to training in the CAN data (particularly for hair and skin colour). Altogether, our observations are consistent with common variation of eye and hair colour having a relatively simple genetic architecture, which is well captured by HIrisPlex-S, even in admixed Latin Americans (with partial European ancestry). By contrast, since skin pigmentation is a more polygenic trait, accuracy is more sensitive to prediction SNP set size, although here this effect was only apparent for a quantitative measure of skin pigmentation. Our results support the use of HIrisPlex-S in the prediction of categorical pigmentation traits for forensic purposes in Latin America, while illustrating the impact of training datasets on its accuracy.

Evaluation of supervised machine-learning methods for predicting appearance traits from DNA

The prediction of human externally visible characteristics (EVCs) based solely on DNA information has become an established approach in forensic and anthropological genetics in recent years. While for a large set of EVCs, predictive models have already been established using multinomial logistic regression (MLR), the prediction performances of other possible classification methods have not been thoroughly investigated thus far. Motivated by the question to identify a potential classifier that outperforms these specific trait models, we conducted a systematic comparison between the widely used MLR and three popular machine learning (ML) classifiers, namely support vector machines (SVM), random forest (RF) and artificial neural networks (ANN), that have shown good performance outside EVC prediction. As examples, we used eye, hair and skin color categories as phenotypes and genotypes based on the previously established IrisPlex, HIrisPlex, and HIrisPlex-S DNA markers. We compared and assessed the performances of each of the four methods, complemented by detailed hyperparameter tuning that was applied to some of the methods in order to maximize their performance. Overall, we observed that all four classification methods showed rather similar performance, with no method being substantially superior to the others for any of the traits, although performances varied slightly across the different traits and more so across the trait categories. Hence, based on our findings, none of the ML methods applied here provide any advantage on appearance prediction, at least when it comes to the categorical pigmentation traits and the selected DNA markers used here.

Genome-wide association study in almost 195,000 individuals identifies 50 previously unidentified genetic loci for eye color

Human eye color is highly heritable, but its genetic architecture is not yet fully understood. We report the results of the largest genome-wide association study for eye color to date, involving up to 192,986 European participants from 10 populations. We identify 124 independent associations arising from 61 discrete genomic regions, including 50 previously unidentified. We find evidence for genes involved in melanin pigmentation, but we also find associations with genes involved in iris morphology and structure. Further analyses in 1636 Asian participants from two populations suggest that iris pigmentation variation in Asians is genetically similar to Europeans, albeit with smaller effect sizes. Our findings collectively explain 53.2% (95% confidence interval, 45.4 to 61.0%) of eye color variation using common single-nucleotide polymorphisms. Overall, our study outcomes demonstrate that the genetic complexity of human eye color considerably exceeds previous knowledge and expectations, highlighting eye color as a genetically highly complex human trait.

The color of normal: How a Eurocentric focus erases pigmentation complexity

OBJECTIVES

Skin pigmentation is both a highly variable and highly visible human phenotypic trait. Investigations into the biology and origins of this variation have been the focus of research in the fields of dermatology, anthropology, and forensic science, among others. This manuscript explores how much of what we know about the biology, genetics, and evolutionary origins of pigmentation has been strongly influenced by investigations and applications that focus on lighter skin.

METHODS

I reviewed literature from the fields of dermatology, anthropology and evolutionary genetics, and forensic science to assess how perceptions of lighter skin as the "normal" state in humans can shape the ways that knowledge is gathered and applied in these fields.

RESULTS

This normalization of lighter skin has impacted common tools used in dermatology and shaped the framework of dermatological education. A strong Eurocentric bias has shaped our understanding of the genetic architecture of pigmentary traits, which influences the ways in we understand the evolutionary processes leading to modern pigmentation diversity. Finally, I discuss how these biases in pigmentation genetics work in combination with phenotypic systems that privilege predicting lighter pigmentation variation to impede accurate prediction of intermediate phenotypes, particularly in individuals with ancestry from multiple populations. This can lead to a disproportionate targeting of already over-policed populations with darker skin.

CONCLUSIONS

Potential changes to how we conceptualize clinical and basic pigmentation research may help to reduce existing health disparities and improve understanding of pigmentation genetic architecture and how this knowledge is applied in forensic contexts.

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

As the field of forensic DNA analysis has started to transition from genetics to genomics, new methods to aid in crime scene investigations have arisen. The development of informative single nucleotide polymorphism (SNP) markers has led the forensic community to question if DNA can be a reliable "eye-witness" and whether the data it provides can shed light on unknown perpetrators. We have developed an assay called the Ion AmpliSeq™ PhenoTrivium Panel, which combines three groups of markers: 41 phenotype- and 163 ancestry-informative autosomal SNPs together with 120 lineage-specific Y-SNPs. Here, we report the results of testing the assay's sensitivity and the predictions obtained for known reference samples. Moreover, we present the outcome of a blind study performed on real casework samples in order to understand the value and reliability of the information that would be provided to police investigators. Furthermore, we evaluated the accuracy of admixture prediction in Converge™ Software. The results show the panel to be a robust and sensitive assay which can be used to analyze casework samples. We conclude that the combination of the obtained predictions of phenotype, biogeographical ancestry, and male lineage can serve as a potential lead in challenging police investigations such as cold cases or cases with no suspect.

Transforming the study of organisms: Phenomic data models and knowledge bases

The rapidly decreasing cost of gene sequencing has resulted in a deluge of genomic data from across the tree of life; however, outside a few model organism databases, genomic data are limited in their scientific impact because they are not accompanied by computable phenomic data. The majority of phenomic data are contained in countless small, heterogeneous phenotypic data sets that are very difficult or impossible to integrate at scale because of variable formats, lack of digitization, and linguistic problems. One powerful solution is to represent phenotypic data using data models with precise, computable semantics, but adoption of semantic standards for representing phenotypic data has been slow, especially in biodiversity and ecology. Some phenotypic and trait data are available in a semantic language from knowledge bases, but these are often not interoperable. In this review, we will compare and contrast existing ontology and data models, focusing on nonhuman phenotypes and traits. We discuss barriers to integration of phenotypic data and make recommendations for developing an operationally useful, semantically interoperable phenotypic data ecosystem.

Iris Colour and the Risk of Developing Uveal Melanoma

Uveal melanoma (UM) is a global disease which especially occurs in elderly people. Its incidence varies widely between populations, with the highest incidence among Caucasians, and a South-to-North increase in Europe. As northern Europeans often have blond hair and light eyes, we wondered whether iris colour may be a predisposing factor for UM and if so, why. We compared the distribution of iris colour between Dutch UM patients and healthy Dutch controls, using data from the Rotterdam Study (RS), and reviewed the literature regarding iris colour. We describe molecular mechanisms that might explain the observed associations. When comparing a group of Dutch UM patients with controls, we observed that individuals from Caucasian ancestry with a green/hazel iris colour (Odds Ratio (OR) = 3.64, 95% Confidence Interval (CI) 2.57-5.14) and individuals with a blue/grey iris colour (OR = 1.38, 95% CI 1.04-1.82) had a significantly higher crude risk of UM than those with brown eyes. According to the literature, this may be due to a difference in the function of pheomelanin (associated with a light iris colour) and eumelanin (associated with a brown iris colour). The combination of light-induced stress and aging may affect pheomelanin-carrying melanocytes in a different way than eumelanin-carrying melanocytes, increasing the risk of developing a malignancy.

The impact of correlations between pigmentation phenotypes and underlying genotypes on genetic prediction of pigmentation traits

Predicting appearance phenotypes from genotypes is relevant for various areas of human genetic research and applications such as genetic epidemiology, human history, anthropology, and particularly in forensics. Many appearance phenotypes, and thus their underlying genotypes, are highly correlated, with pigmentation traits serving as primary examples. However, all available genetic prediction models, including those for pigmentation traits currently used in forensic DNA phenotyping, ignore phenotype correlations. Here, we investigated the impact of appearance phenotype correlations on genetic appearance prediction in the exemplary case of three pigmentation traits. We used data for categorical eye, hair and skin colour as well as 41 DNA markers utilized in the recently established HIrisPlex-S system from 762 individuals with complete phenotype and genotype information. Based on these data, we performed genetic prediction modelling of eye, hair and skin colour via three different strategies, namely the established approach of predicting phenotypes solely based on genotypes while not considering phenotype correlations, and two novel approaches that considered phenotype correlations, either incorporating truly observed correlated phenotypes or DNA-predicted correlated phenotypes in addition to the DNA predictors. We found that using truly observed correlated pigmentation phenotypes as additional predictors increased the DNA-based prediction accuracies for almost all eye, hair and skin colour categories, with the largest increase for intermediate eye colour, brown hair colour, dark to black skin colour, and particularly for dark skin colour. Outcomes of dedicated computer simulations suggest that this prediction accuracy increase is due to the additional genetic information that is implicitly provided by the truly observed correlated pigmentation phenotypes used, yet not covered by the DNA predictors applied. In contrast, considering DNA-predicted correlated pigmentation phenotypes as additional predictors did not improve the performance of the genetic prediction of eye, hair and skin colour, which was in line with the results from our computer simulations. Hence, in practical applications of DNA-based appearance prediction where no phenotype knowledge is available, such as in forensic DNA phenotyping, it is not advised to use DNA-predicted correlated phenotypes as predictors in addition to the DNA predictors. In the very least, this is not recommended for the pigmentation traits and the established pigmentation DNA predictors tested here.

Association between brown eye colour in rs12913832:GG individuals and SNPs in TYR, TYRP1, and SLC24A4

The genotype of a single SNP, rs12913832, is the primary predictor of blue and brown eye colours. The genotypes rs12913832:AA and rs12913832:GA are most often observed in individuals with brown eye colours, whereas rs12913832:GG is most often observed in individuals with blue eye colours. However, approximately 3% of Europeans with the rs12913832:GG genotype have brown eye colours. The purpose of the study presented here was to identify variants that explain brown eye colour formation in individuals with the rs12913832:GG genotype. Genes and regulatory regions surrounding SLC24A4, TYRP1, SLC24A5, IRF4, TYR, and SLC45A2, as well as the upstream region of OCA2 within the HERC2 gene were sequenced in a study comprising 40 individuals with the rs12913832:GG genotype. Of these, 24 individuals were considered to have blue eye colours and 16 individuals were considered to have brown eye colours. We identified 211 variants within the SLC24A4, TYRP1, IRF4, and TYR target regions associated with eye colour. Based on in silico analyses of predicted variant effects we recognized four variants, TYRP1 rs35866166:C, TYRP1 rs62538956:C, SLC24A4 rs1289469:C, and TYR rs1126809:G, to be the most promising candidates for explanation of brown eye colour in individuals with the rs12913832:GG genotype. Of the 16 individuals with brown eye colours, 14 individuals had four alleles, whereas the alleles were rare in the blue eyed individuals. rs35866166, rs62538956, and rs1289469 were for the first time found to be associated with pigmentary traits, whilst rs1126809 was previously found to be associated with pigmentary variation. To improve prediction of eye colours we suggest that future eye colour prediction models should include rs35866166, rs62538956, rs1289469, and rs1126809.

Optimizing the genetic prediction of the eye and hair color for North Eurasian populations

Background

Predicting the eye and hair color from genotype became an established and widely used tool in forensic genetics, as well as in studies of ancient human populations. However, the accuracy of this tool has been verified on the West and Central Europeans only, while populations from border regions between Europe and Asia (like Caucasus and Ural) also carry the light pigmentation phenotypes.

Results

We phenotyped 286 samples collected across North Eurasia, genotyped them by the standard HIrisPlex-S markers and found that predictive power in Caucasus/Ural/West Siberian populations is reasonable but lower than that in West Europeans. As these populations have genetic ancestries different from that of West Europeans, we hypothesized they may carry a somewhat different allele spectrum. Thus, for all samples we performed the exome sequencing additionally enriched with the 53 genes and intergenic regions known to be associated with the eye/hair color. Our association analysis replicated the importance of the key previously known SNPs but also identified five new markers whose eye color prediction power for the studied populations is compatible with the two major previously well-known SNPs. Four out of these five SNPs lie within the HERС2 gene and the fifth in the intergenic region. These SNPs are found at high frequencies in most studied populations. The released dataset of exomes from Russian populations can be further used for population genetic and medical genetic studies.

Conclusions

This study demonstrated that precision of the established systems for eye/hair color prediction from a genotype is slightly lower for the populations from the border regions between Europe and Asia that for the West Europeans. However, this precision can be improved if some newly revealed predictive SNPs are added into the panel. We discuss that the replication of these pigmentation-associated SNPs on the independent North Eurasian sample is needed in the future studies.

Eye color prediction using the IrisPlex system: a limited pilot study in the Iraqi population

Background

Forensic DNA phenotyping has gained momentum in the recent past due to the prediction of externally visible characters (EVCs) from the biological sample. The most common phenotypes like eye, hair, and skin color are predicted from the biological samples using a web-based system called IrisPlex. Based on six genetic SNPs, the IrisPlex system is developed and validated for its prediction accuracy in diverse ethnic groups worldwide. In previous studies, this system proved to have significant prediction accuracy. The EVCs vary substantially based on different geographical locations. Hence, the objective of this study was to validate the accuracy of the IrisPlex system in predicting the eye colors in the Iraqi population.

Methods

Six genetic single-nucleotide polymorphisms SNPs (HERC2-rs12913832, OCA2- rs1800407, SLC24A4-rs12896399, SLC45A2- rs16891982, TYR-rs1393350, and IRF4- rs12203592) in 58 Iraqi subjects were performed using Sequenom MassARRAY Genotyping. According to Liu et al., a predicted probability of 0.7 was considered as the threshold.

Results

Participants in this study of brown color were observed in 44.83%, intermediate in 43.1%, and blue in 12.07%. Completely predictive accuracy is obtained in 1; we observed the AUC at threshold 0.7 was 0.91 for brown, 0.79 for blue, and 0.60 for intermediate eye color. The sensitivity was 42.85% for blue, 0% for intermediate eye color, and 100% for brown-colored eye. Specificity was 100% for blue, 100% for intermediate, and 78.13% for brown eye color.

Conclusion

Hence, it was concluded that the prediction accuracy of the IrisPlex system for blue and brown color eye in the Iraqi population is significant in the studied population size. However, a pivotal study with larger sample size is required to represent the prediction accuracy of the IrisPlex system in the whole Iraqi population.

The Use of Forensic DNA Phenotyping in Predicting Appearance and Biogeographic Ancestry

BACKGROUND

Persons whose identifying DNA profile (STR profile) is not yet known to the ingvestigating authorities cannot be identified by standard forensic DNA analysis (STR profiling) as it is now practiced. In view of the current public debate, particularly in Germany, on the legalization of so-called forensic DNA phenotyping, we present its scientific basis, societal aspects, and forensic applications and describe the analytic techniques that are now available.

METHODS

This review is based on pertinent publications that were retrieved by a selective search in PubMed and in public media, and on the authors' own research.

RESULTS

Forensically validated DNA test systems are available for the categorization of eye, hair, and skin color and the inference of continental biogeographic ancestry. As for statistical measures of test accuracy, the AUC (area under the curve) values lie in the range 0.74-0.99 for eye color, 0.64-0.94 for hair color, and 0.72-0.99 for skin color, depending on the predictive model and color category used.The corre- sponding positive predictive values (PPV) are lower. Empirical social-scientific research on forensic DNA phenotyping has shown that preserving privacy and protecting against discrimination are major ethical and regulatory considerations.

CONCLUSION

All three methods of forensic DNA phenotyping-the predition of exter- nally visible characteristics, biogeographic ancestry, and the estimation of age from crime scene DNA-require a proper regulatory framework and should be used in conjunction with each other. Before forensic DNA phenotyping can be implemented in forensic practice, steps must be taken to minimize the risks of violation of privacy scrimination and to ensure that these methods are used transpar- ently and proportionately.

Skin pigmentation and genetic variants in an admixed Brazilian population of primarily European ancestry

Although many genes have been shown to be associated with human pigmentary traits and forensic prediction assays exist (e.g. HIrisPlex-S), the genetic knowledge about skin colour remains incomplete. The highly admixed Brazilian population is an interesting study population for investigation of the complex genotype-phenotype architecture of human skin colour because of its large variation. Here, we compared variants in 22 pigmentary genes with quantitative skin pigmentation levels on the buttock, arm, and forehead areas of 266 genetically admixed Brazilian individuals. The genetic ancestry of each individual was estimated by typing 46 AIM-InDels. The mean proportion of genetic ancestry was 68.8% European, 20.8% Sub-Saharan African, and 10.4% Native American. A high correlation (adjusted R² = 0.65, p < 0.05) was observed between nine SNPs and quantitative skin pigmentation using multiple linear regression analysis. The correlations were notably smaller between skin pigmentation and biogeographic ancestry (adjusted R² = 0.45, p < 0.05), or markers in the leading forensic skin colour prediction system, the HIrisPlex-S (adjusted R² = 0.54, p < 0.05). Four of the nine SNPs, OCA2 rs1448484 (rank 2), APBA2 rs4424881 (rank 4), MFSD12 rs10424065 (rank 8), and TYRP1 1408799 (rank 9) were not investigated as part of the HIrisPlex-S selection process, and therefore not included in the HIrisPlex-S model. Our results indicate that these SNPs account for a substantial part of the skin colour variation in individuals of admixed ancestry. Hence, we suggest that these SNPs are considered when developing future skin colour prediction models.

The challenge of predicting human pigmentation traits in degraded bone samples with the MPS-based HIrisPlex-S system

Identification of human remains is an important part of human DNA analysis studies. STR and mitochondrial DNA markers are well suited for the analysis of degraded biological samples including bone material. However, these DNA markers may be useless when reference material is not available. In these cases, predictive DNA analysis can support the process of human identification by providing investigative leads. Forensic DNA phenotyping has progressed significantly by offering new methods based on massively parallel sequencing technology, but the frequent degradation processes observed in skeletal remains can make analysis of such samples challenging. In this study, we demonstrate the usefulness of a recently established Ion AmpliSeq^TM HIrisPlex-S panel using Ion Torrent technology for analyzing bone samples that show different levels of DNA degradation. In total, 63 bone samples at post-mortem intervals up to almost 80 years were genotyped and eye, hair and skin colour predictions were performed using the HIrisPlex-S models. Following the recommended coverage thresholds, it was possible to establish full DNA profiles comprising of 41 DNA variants for 35 samples (55.6%). For 5 samples (7.9%) no DNA profiles were generated. The remaining 23 samples (36.5%) produced partial profiles and showed a clear underperformance of 3 HIrisPlex-S SNPs - rs1545397 (OCA2), rs1470608 (OCA2) and rs10756819 (BNC2), all used for skin colour prediction only. None of the 23 samples gave complete genotypes needed for skin colour prediction was obtained, and in 7 of them (25.9%) the 3 underperformed SNPs were the cause. At the same time, the prediction of eye and hair colour using complete IrisPlex and HIrisPlex profiles could be made for these 23 samples in 20 (87.0%) and 12 cases (52.2%), respectively. Complete HIrisPlex-S profiles were generated from as little as 49 pg of template DNA. Five samples for which the HIrisPlex-S analysis failed, consistently failed in standard STR analysis. Importantly, the 3 underperforming SNPs produced significantly lower number of reads in good quality samples. Nonetheless, the AUC loss resulting from missing data for these 3 SNPs is not considered large (≤0.004) and the prediction of pigmentation from partial profiles is also available in the current HPS tool. The study shows that DNA degradation and the resulting loss of data are the most serious challenge to DNA phenotyping of skeletal remains. Although the newly developed HIrisPlex-S panel has been successfully validated in the current research, primer redesign for the 3 underperforming SNPs in the MPS design should be considered in the future.