Exploring legal age estimation using DNA methylation.

Minors (subjects under the legal age, established at this study at 18 years) benefit from a series of legal rights created to protect them and guarantee their welfare. However, throughout the world there are many minors who have no way to prove they are underaged, leading to a great interest in predicting legal age with the highest possible accuracy. Current methods, mainly involving X-ray analysis, are highly invasive, so new methods to predict legal age are being studied, such as DNA methylation. To further such studies, we created two age prediction models based on five epigenetic markers: cg21572722 (ELOVL2), cg02228185 (ASPA), cg06639320 (FHL2), cg19283806 (CCDC102B) and cg07082267 (no associated gene), that were analysed in blood samples to determine possible limitations regarding DNA methylation as an effective tool for legal age estimation. A wide age range prediction model was created using a broad set of samples (14-94 years) yielding a mean absolute error (MAE) of ±4.32 years. A second model, the constrained age prediction model, was created using a reduced range of samples (14-25 years) yielding an MAE of ±1.54 years. Both models, in addition to Horvath's Skin & Blood epigenetic clock, were evaluated using a test set comprising 732 pairs of 18-year-old twins (N=426 monozygotic (MZ) and N=306 dizygotic (DZ) pairs), representing a relevant age of study. Through analysis of the two former age prediction models, we found that constraining the age of the samples forming the training set around the desired age of study significantly reduced the prediction error (from MAE: ±4.07 and ±4.27 years for MZ and DZ twins, respectively; to ±1.31 and ±1.3 years). However, despite low prediction errors, DNA methylation models are still prone to classify same-aged individuals in different categories (minors or adults), despite each sample belonging to the same twin pair. Additional evaluation of Horvath's Skin & Blood model (391 CpGs) led to similar results in terms of age prediction errors than if using only five epigenetic markers (MAE: ±1.87 and ±1.99 years for MZ and DZ twins, respectively).

Inference of tobacco and alcohol consumption habits from DNA methylation analysis of blood.

DNA methylation has become a biomarker of great interest in the forensic and clinical fields. In criminal investigations, the study of this epigenetic marker has allowed the development of DNA intelligence tools providing information that can be useful for investigators, such as age prediction. Following a similar trend, when the origin of a sample in a criminal scenario is unknown, the inference of an individual's lifestyle such as tobacco use and alcohol consumption could provide relevant information to help in the identification of DNA donors at the crime scene. At the same time, in the clinical domain, prediction of these trends of consumption could allow the identification of people at risk or better identification of the causes of different pathologies. In the present study, DNA methylation data from the UK AIRWAVE study was used to build two binomial logistic models for the inference of smoking and drinking status. A total of 348 individuals (116 non-smokers, 116 former smokers and 116 smokers) plus a total of 237 individuals (79 non-drinkers, 79 moderate drinkers and 79 drinkers) were used for development of tobacco and alcohol consumption prediction models, respectively. The tobacco prediction model was composed of two CpGs (cg05575921 in AHRR and cg01940273) and the alcohol prediction model three CpGs (cg06690548 in SLC7A11, cg0886875 and cg21294714 in MIR4435-2HG), providing correct classifications of 86.49% and 74.26%, respectively. Validation of the models was performed using leave-one-out cross-validation. Additionally, two independent testing sets were also assessed for tobacco and alcohol consumption. Considering that the consumption of these substances could underlie accelerated epigenetic ageing patterns, the effect of these lifestyles on the prediction of age was evaluated. To do that, a quantile regression model based on previous studies was generated, and the potential effect of tobacco and alcohol consumption with the epigenetic age was assessed. The Wilcoxon test was used to evaluate the residuals generated by the model and no significant differences were observed between the categories analyzed.

Development of an epigenetic age predictor for costal cartilage with a simultaneous somatic tissue differentiation system.

Age prediction from DNA has been a topic of interest in recent years due to the promising results obtained when using epigenetic markers. Since DNA methylation gradually changes across the individual's lifetime, prediction models have been developed accordingly for age estimation. The tissue-dependence for this biomarker usually necessitates the development of tissue-specific age prediction models, in this way, multiple models for age inference have been constructed for the most commonly encountered forensic tissues (blood, oral mucosa, semen). The analysis of skeletal remains has also been attempted and prediction models for bone have now been reported. Recently, the VISAGE Enhanced Tool was developed for the simultaneous DNA methylation analysis of 8 age-correlated loci using targeted high-throughput sequencing. It has been shown that this method is compatible with epigenetic age estimation models for blood, buccal cells, and bone. Since when dealing with decomposed cadavers or postmortem samples, cartilage samples are also an important biological source, an age prediction model for cartilage has been generated in the present study based on methylation data collected using the VISAGE Enhanced Tool. In this way, we have developed a forensic cartilage age prediction model using a training set composed of 109 samples (19-74 age range) based on DNA methylation levels from three CpGs in FHL2, TRIM59 and KLF14, using multivariate quantile regression which provides a mean absolute error (MAE) of ± 4.41 years. An independent testing set composed of 72 samples (19-75 age range) was also analyzed and provided an MAE of ± 4.26 years. In addition, we demonstrate that the 8 VISAGE markers, comprising EDARADD, TRIM59, ELOVL2, MIR29B2CHG, PDE4C, ASPA, FHL2 and KLF14, can be used as tissue prediction markers which provide reliable blood, buccal cells, bone, and cartilage differentiation using a developed multinomial logistic regression model. A training set composed of 392 samples (n = 87 blood, n = 86 buccal cells, n = 110 bone and n = 109 cartilage) was used for building the model (correct classifications: 98.72%, sensitivity: 0.988, specificity: 0.996) and validation was performed using a testing set composed of 192 samples (n = 38 blood, n = 36 buccal cells, n = 46 bone and n = 72 cartilage) showing similar predictive success to the training set (correct classifications: 97.4%, sensitivity: 0.968, specificity: 0.991). By developing both a new cartilage age model and a tissue differentiation model, our study significantly expands the use of the VISAGE Enhanced Tool while increasing the amount of DNA methylation-based information obtained from a single sample and a single forensic laboratory analysis. Both models have been placed in the open-access Snipper forensic classification website.

Epigenetic age estimation in saliva and in buccal cells.

Age estimation based on epigenetic markers is a DNA intelligence tool with the potential to provide relevant information for criminal investigations, as well as to improve the inference of age-dependent physical characteristics such as male pattern baldness or hair color. Age prediction models have been developed based on different tissues, including saliva and buccal cells, which show different methylation patterns as they are composed of different cell populations. On many occasions in a criminal investigation, the origin of a sample or the proportion of tissues is not known with certainty, for example the provenance of cigarette butts, so use of combined models can provide lower prediction errors. In the present study, two tissue-specific and seven age-correlated CpG sites were selected from publicly available data from the Illumina HumanMethylation 450 BeadChip and bibliographic searches, to help build a tissue-dependent, and an age-prediction model, respectively. For the development of both models, a total of 184 samples (N = 91 saliva and N = 93 buccal cells) ranging from 21 to 86 years old were used. Validation of the models was performed using either k-fold cross-validation and an additional set of 184 samples (N = 93 saliva and N = 91 buccal cells, 21-86 years old). The tissue prediction model was developed using two CpG sites (HUNK and RUNX1) based on logistic regression that produced a correct classification rate for saliva and buccal swab samples of 88.59 % for the training set, and 83.69 % for the testing set. Despite these high success rates, a combined age prediction model was developed covering both saliva and buccal cells, using seven CpG sites (cg10501210, LHFPL4, ELOVL2, PDE4C, HOXC4, OTUD7A and EDARADD) based on multivariate quantile regression giving a median absolute error (MAE): ± 3.54 years and a correct classification rate ( %CP±PI) of 76.08 % for the training set, and an MAE of ± 3.66 years and a %CP±PI of 71.19 % for the testing set. The addition of tissue-of origin as a co-variate to the model was assessed, but no improvement was detected in age predictions. Finally, considering the limitations usually faced by forensic DNA analyses, the robustness of the model and the minimum recommended amount of input DNA for bisulfite conversion were evaluated, considering up to 10 ng of genomic DNA for reproducible results. The final multivariate quantile regression age predictor based on the models we developed has been placed in the open-access Snipper forensic classification website.

A common epigenetic clock from childhood to old age.

Forensic age estimation is a DNA intelligence tool that forms an important part of Forensic DNA Phenotyping. Criminal cases with no suspects or with unsuccessful matches in searches on DNA databases; human identification analyses in mass disasters; anthropological studies or legal disputes; all benefit from age estimation to gain investigative leads. Several age prediction models have been developed to date based on DNA methylation. Although different DNA methylation technologies as well as diverse statistical methods have been proposed, most of them are based on blood samples and mainly restricted to adult age ranges. In the current study, we present an extended age prediction model based on 895 evenly distributed Spanish DNA blood samples from 2 to 104 years old. DNA methylation levels were detected using Agena Bioscience EpiTYPER® technology for a total of seven CpG sites located at seven genomic regions: ELOVL2, ASPA, PDE4C, FHL2, CCDC102B, MIR29B2CHG and chr16:85395429 (GRCh38). The accuracy of the age prediction system was tested by comparing three statistical methods: quantile regression (QR), quantile regression neural network (QRNN) and quantile regression support vector machine (QRSVM). The most accurate predictions were obtained when using QRNN or QRSVM (mean absolute prediction error, MAE of ± 3.36 and ± 3.41, respectively). Validation of the models with an independent Spanish testing set (N = 152) provided similar accuracies for both methods (MAE: ± 3.32 and ± 3.45, respectively). The main advantage of using quantile regression statistical tools lies in obtaining age-dependent prediction intervals, fitting the error to the estimated age. An additional analysis of dimensionality reduction shows a direct correlation of increased error and a reduction of correct classifications as the training sample size is reduced. Results indicated that a minimum sample size of six samples per year-of-age covered by the training set is recommended to efficiently capture the most inter-individual variability..

A Comparison of Forensic Age Prediction Models Using Data From Four DNA Methylation Technologies.

Individual age estimation can be applied to criminal, legal, and anthropological investigations. DNA methylation has been established as the biomarker of choice for age prediction, since it was observed that specific CpG positions in the genome show systematic changes during an individual's lifetime, with progressive increases or decreases in methylation levels. Subsequently, several forensic age prediction models have been reported, providing average age prediction error ranges of ±3-4 years, using a broad spectrum of technologies and underlying statistical analyses. DNA methylation assessment is not categorical but quantitative. Therefore, the detection platform used plays a pivotal role, since quantitative and semi-quantitative technologies could potentially result in differences in detected DNA methylation levels. In the present study, we analyzed as a shared sample pool, 84 blood-based DNA controls ranging from 18 to 99 years old using four different technologies: EpiTYPER®, pyrosequencing, MiSeq, and SNaPshotTM. The DNA methylation levels detected for CpG sites from ELOVL2, FHL2, and MIR29B2 with each system were compared. A restricted three CpG-site age prediction model was rebuilt for each system, as well as for a combination of technologies, based on previous training datasets, and age predictions were calculated accordingly for all the samples detected with the previous technologies. While the DNA methylation patterns and subsequent age predictions from EpiTYPER®, pyrosequencing, and MiSeq systems are largely comparable for the CpG sites studied, SNaPshotTM gives bigger differences reflected in higher predictive errors. However, these differences can be reduced by applying a z-score data transformation.

Tracking age-correlated DNA methylation markers in the young.

DNA methylation is the most extensively studied epigenetic signature, with a large number of studies reporting age-correlated CpG sites in overlapping genes. However, most of these studies lack sample coverage of individuals under 18 years old and therefore little is known about the progression of DNA methylation patterns in children and adolescents. In the present study we aimed to select candidate age-correlated DNA methylation markers based on public datasets from Illumina BeadChip arrays and previous publications, then to explore the resulting markers in 209 blood samples from donors aged between 2 to 18 years old using the EpiTYPER® DNA methylation analysis system. Results from our analyses identified six genes highly correlated with age in the young, in particular the gene KCNAB3, which indicates its potential as a highly informative and specific age biomarker for childhood and adolescence. We outline a preliminary age prediction model based on quantile regression that uses data from the six CpG sites most strongly correlated with age ranges extended to include children and adolescents.

Comparing student and staff perceptions of the "Educational Climate" in Spanish Dental Schools using the Dundee Ready Education Environment Measure.

OBJECTIVE: To compare the perceptions of students and teachers of the "Educational Climate" (EC) in Spanish public dental schools. METHODS: A group of 1064 students and 354 teachers from six Spanish public dental schools responded to the DREEM questionnaire. This has 50 items grouped into five subscales: perception of learning (Learning); perception of teachers (Teachers); academic self-perceptions (Academic); perception of the atmosphere in the faculty (Atmosphere); and social self-perceptions (Social). The DREEM scale provides results for each item, each subscale and the overall EC. RESULTS: The EC scores were 123.2 (61.6%) for the students and 134.1 (67.0%) for the teachers (P<.001). The scores of the students and teachers for the subscales were, respectively: 27.9 (58.1%) and 30.2 (63.0 %) for Learning (P<.001); 26.8 (60.9%) and 32.6 (74.1%) for Teachers (P<.001); 20.7 (64.7%) and 20.5 (64.0%) for Academic (P=.333); 29.9 (62.3%) and 33.7 (70.3%) for Atmosphere (P<.001); and 17.9 (64.0%) and 16.9 (60.5%) for Social (P<.001). The students identified six problematic items (12.0 %) compared to only two (4.0 %) highlighted by the teachers. CONCLUSION: The students and teachers considered the EC to be "more positive than negative" in Spanish public dental schools; and the different subscales to be "positive and acceptable." The teachers did, however, evaluate the EC, and specifically the learning-teaching process, more positively than their students, identifying fewer problematic educational aspects. Both groups agreed on the need to: improve support systems for students who suffer from stress and reduce teaching based on "factual learning."

Cytokine-based Predictive Models to Estimate the Probability of Chronic Periodontitis: Development of Diagnostic Nomograms.

Although a distinct cytokine profile has been described in the gingival crevicular fluid (GCF) of patients with chronic periodontitis, there is no evidence of GCF cytokine-based predictive models being used to diagnose the disease. Our objectives were: to obtain GCF cytokine-based predictive models; and develop nomograms derived from them. A sample of 150 participants was recruited: 75 periodontally healthy controls and 75 subjects affected by chronic periodontitis. Sixteen mediators were measured in GCF using the Luminex 100™ instrument: GMCSF, IFNgamma, IL1alpha, IL1beta, IL2, IL3, IL4, IL5, IL6, IL10, IL12p40, IL12p70, IL13, IL17A, IL17F and TNFalpha. Cytokine-based models were obtained using multivariate binary logistic regression. Models were selected for their ability to predict chronic periodontitis, considering the different role of the cytokines involved in the inflammatory process. The outstanding predictive accuracy of the resulting smoking-adjusted models showed that IL1alpha, IL1beta and IL17A in GCF are very good biomarkers for distinguishing patients with chronic periodontitis from periodontally healthy individuals. The predictive ability of these pro-inflammatory cytokines was increased by incorporating IFN gamma and IL10. The nomograms revealed the amount of periodontitis-associated imbalances between these cytokines with pro-inflammatory and anti-inflammatory effects in terms of a particular probability of having chronic periodontitis.

Development of a methylation marker set for forensic age estimation using analysis of public methylation data and the Agena Bioscience EpiTYPER system.

Individual age estimation has the potential to provide key information that could enhance and extend DNA intelligence tools. Following predictive tests for externally visible characteristics developed in recent years, prediction of age could guide police investigations and improve the assessment of age-related phenotype expression patterns such as hair colour changes and early onset of male pattern baldness. DNA methylation at CpG positions has emerged as the most promising DNA tests to ascertain the individual age of the donor of a biological contact trace. Although different methodologies are available to detect DNA methylation, EpiTYPER technology (Agena Bioscience, formerly Sequenom) provides useful characteristics that can be applied as a discovery tool in localized regions of the genome. In our study, a total of twenty-two candidate genomic regions, selected from the assessment of publically available data from the Illumina HumanMethylation 450 BeadChip, had a total of 177 CpG sites with informative methylation patterns that were subsequently investigated in detail. From the methylation analyses made, a novel age prediction model based on a multivariate quantile regression analysis was built using the seven highest age-correlated loci of ELOVL2, ASPA, PDE4C, FHL2, CCDC102B, C1orf132 and chr16:85395429. The detected methylation levels in these loci provide a median absolute age prediction error of ±3.07years and a percentage of prediction error relative to the age of 6.3%. We report the predictive performance of the developed model using cross validation of a carefully age-graded training set of 725 European individuals and a test set of 52 monozygotic twin pairs. The multivariate quantile regression age predictor, using the CpG sites selected in this study, has been placed in the open-access Snipper forensic classification website.

Psychometric validation of the Spanish version of the Dundee Ready Education Environment Measure applied to dental students.

AIM: To carry out a psychometric evaluation of the Spanish-language version of the Dundee Ready Education Environment Measure (DREEM) applied to dental students. METHODS: A total of 1,391 students from nine Spanish public schools of dentistry responded to the DREEM questionnaire. To analyse the reliability of the DREEM questionnaire, the internal consistency was assessed and a 'test-retest' carried out. Validity was evaluated through analysis of item response rate, floor and ceiling effects, corrected item-total and item-subscale correlations and factor structure. A confirmatory factor analysis was performed to analyse the structure of the original DREEM scale. RESULTS: Cronbach's alpha coefficient for the 'Educational Climate' (EC) global scale was 0.92. In the subscales, the 'observed' Cronbach's alpha coefficients ranged between 0.57 and 0.79 and were higher than the 'expected' ones; except for the Social subscale. In the DREEM questionnaire, all of the corrected correlation coefficients between the items and the EC global scale, and the items and their corresponding subscales, were >0.2; except for items 50 and 17. All goodness-of-fit indices of confirmatory factor analysis showed acceptable values (close to one or zero, depending on the case), and there was consistency in the results. CONCLUSIONS: The Spanish-language version of the DREEM questionnaire is a reliable and valid instrument for analysing the EC for dental students and its factor structure is supported by the data. Although our findings indicate that the DREEM may be as culturally independent as was originally stated, more research should be directed at verifying the factor structure in various languages and cultural environments.

Further development of forensic eye color predictive tests.

In forensic analysis predictive tests for external visible characteristics (or EVCs), including inference of iris color, represent a potentially useful tool to guide criminal investigations. Two recent studies, both focused on forensic testing, have analyzed single nucleotide polymorphism (SNP) genotypes underlying common eye color variation (Mengel-From et al., Forensic Sci. Int. Genet. 4:323 and Walsh et al., Forensic Sci. Int. Genet. 5:170). Each study arrived at different recommendations for eye color predictive tests aiming to type the most closely associated SNPs, although both confirmed rs12913832 in HERC2 as the key predictor, widely recognized as the most strongly associated marker with blue and brown iris colors. Differences between these two studies in identification of other eye color predictors may partly arise from varying approaches to assigning phenotypes, notably those not unequivocally blue or dark brown and therefore occupying an intermediate iris color continuum. We have developed two single base extension assays typing 37 SNPs in pigmentation-associated genes to study SNP-genotype based prediction of eye, skin, and hair color variation. These assays were used to test the performance of different sets of eye color predictors in 416 subjects from six populations of north and south Europe. The presence of a complex and continuous range of intermediate phenotypes distinct from blue and brown eye colors was confirmed by establishing eye color populations compared to genetic clusters defined using Structure software. Our study explored the effect of an expanded SNP combination beyond six markers has on the ability to predict eye color in a forensic test without extending the SNP assay excessively - thus maintaining a balance between the test's predictive value and an ability to reliably type challenging DNA with a multiplex of manageable size. Our evaluation used AUC analysis (area under the receiver operating characteristic curves) and naïve Bayesian likelihood-based classification approaches. To provide flexibility in SNP-based eye color predictive tests in forensic applications we modified an online Bayesian classifier, originally developed for genetic ancestry analysis, to provide a straightforward system to assign eye color likelihoods from a SNP profile combining additional informative markers from the predictors analyzed by our study plus those of Walsh and Mengel-From. Two advantages of the online classifier is the ability to submit incomplete SNP profiles, a common occurrence when typing challenging DNA, and the ability to handle physically linked SNPs showing independent effect, by allowing the user to input frequencies from SNP pairs or larger combinations. This system was used to include the submission of frequency data for the SNP pair rs12913832 and rs1129038: indicated by our study to be the two SNPs most closely associated to eye color.