Statistical model for degraded DNA samples and adjusted probabilities for allelic drop-out

An exploratory view into allelic drop-out of sequenced autosomal STRs

As massively parallel sequencing is implemented in forensic genetics, an understanding of sequence data must accompany these advancements, that is, accurate modeling of data for proper statistical analysis. Allelic drop-out, a common stochastic effect seen in genetic data, is often modeled in statistical analysis of STR results. This proof-of-concept study sequenced several serial dilutions of a standard sample ranging from 4 ng to 7.82 pg to evaluate allelic drop-out trends on a select panel of autosomal STRs using the ForenSeq™ DNA Signature Prep Kit, Primer Set A on the Illumina MiSeq FGx. Parameters assessed included locus, profile, and run specific information. A majority of the allelic drop-out occurred in DNA concentrations less than 31.25 pg. Statistical results indicated a need for locus-specific modeling based on STR descriptors, like simple versus compound repeat patterns. No correlation was seen between average read count of scored alleles and allelic drop-out at a locus. A statistical correlation was observed between the amount of allelic drop-out and the starting amount of DNA in a sample, average read count of a sample, and total read count generated on a flow cell. This study supports using common allelic drop-out factors used in fragment length analysis on sequenced STRs while including additional locus, sample, and run specific information. Results demonstrate multiple factors that can be considered when developing probability of allelic drop-out models for sequenced autosomal STRs including locus-specific analysis, total read count of a profile, and total read count sequenced on a flow cell.

Using unique molecular identifiers to improve allele calling in low-template mixtures

PCR artifacts are an ever-present challenge in sequencing applications. These artifacts can seriously limit the analysis and interpretation of low-template samples and mixtures, especially with respect to a minor contributor. In medicine, molecular barcoding techniques have been employed to decrease the impact of PCR error and to allow the examination of low-abundance somatic variation. In principle, it should be possible to apply the same techniques to the forensic analysis of mixtures. To that end, several short tandem repeat loci were selected for targeted sequencing, and a bioinformatic pipeline for analyzing the sequence data was developed. The pipeline notes the relevant unique molecular identifiers (UMIs) attached to each read and, using machine learning, filters the noise products out of the set of potential alleles. To evaluate this pipeline, DNA from pairs of individuals were mixed at different ratios (1-1, 1-9) and sequenced with different starting amounts of DNA (10, 1 and 0.1 ng). Naïvely using the information in the molecular barcodes led to increased performance, with the machine learning resulting in an additional benefit. In concrete terms, using the UMI data results in less noise for a given amount of drop out. For instance, if thresholds are selected that filter out a quarter of the true alleles, using read counts accepts 2381 noise alleles and using raw UMI counts accepts 1726 noise alleles, while the machine learning approach only accepts 307.

Joint DNA-based disaster victim identification

We address computational and statistical aspects of DNA-based identification of victims in the aftermath of disasters. Current methods and software for such identification typically consider each victim individually, leading to suboptimal power of identification and potential inconsistencies in the statistical summary of the evidence. We resolve these problems by performing joint identification of all victims, using the complete genetic data set. Individual identification probabilities, conditional on all available information, are derived from the joint solution in the form of posterior pairing probabilities. A closed formula is obtained for the a priori number of possible joint solutions to a given DVI problem. This number increases quickly with the number of victims and missing persons, posing computational challenges for brute force approaches. We address this complexity with a preparatory sequential step aiming to reduce the search space. The examples show that realistic cases are handled efficiently. User-friendly implementations of all methods are provided in the R package dvir, freely available on all platforms.

DNA degradation in human teeth exposed to thermal stress

Human identification from burned remains poses a challenge to forensic laboratories, and DNA profiling is widely used for this purpose. Our aim was to evaluate the effect of temperature on DNA degradation in human teeth. Thirty teeth were exposed to temperatures of 100, 200, or 400 °C for 60 min. DNA was quantified by Real-Time qPCR (Quantifiler Human DNA Quantification Kit) and fluorescence spectroscopy (Qubit 3.0 Fluorometer). DNA degradation was evaluated by using STR markers (AmpFLSTR Identifiler Plus PCR Amplification Kit) to determine the allele and locus dropout, inter-locus balance, and degradation slope (observed (Oa) to expected (Ea) locus peak height ratio against the molecular weight). Most of the genomic DNA was degraded between 100 °C and 200 °C. At 100 °C, locus dropout ratios showed significant differences between the largest loci (FGA, D7S820, D18S51, D16S539, D2S1338 and CSF1PO) and amelogenin. Inter-locus balance values significantly differed between all dye channels except between NED and PET. The dropout ratio between D18S51 (NED) and amelogenin (PET) can be recommended for the evaluation of DNA degradation. The Oa/Ea regression model can predict locus peak heights in DNA degradation (R2 = 0.7881). These findings may be useful to assess the reliability of DNA typing for human identification in teeth subjected to prolonged incineration.

Suitability of specific soft tissue swabs for the forensic identification of highly decomposed bodies

When decomposition of a recovered body is fairly advanced, identification based on common morphologic features is often impossible. In these cases, short tandem repeat (STR) marker genotyping has established itself as a convenient and reliable alternative. However, at very progressed stages of decomposition, postmortem tissue putrefaction processes can decrease DNA yields considerably. Hence, not all types of tissue are equally suitable for successful STR marker-based postmortem identification. Bone or dental material is often analysed in corpses with advanced decompositional changes. However, processing of these materials is very elaborate and time and resource consuming. We have therefore focused on the suitableness of various types of soft tissue swabs, where DNA extraction is easier and faster. By sampling 28 bodies at various stages of decomposition, we evaluated the suitability of different tissues for genotyping at varying degrees of physical decay. This was achieved by a systematic classification of the sampled bodies by morphological scoring and subsequent analysis of multiple tissue swabs of the aortic wall, urinary bladder wall, brain, liver, oral mucosa and skeletal muscle. In summary, we found variable degrees of suitability of different types of soft tissue swabs for DNA-based identification. Swabs of the aortic wall, the urinary bladder wall and brain tissue yielded the best results - in descending order - even at advanced levels of decay.

Modeling allelic analyte signals for aSTRs in NGS DNA profiles

We describe an adaption of Bright et al.'s work modeling peak height variability in CE-DNA profiles to the modeling of allelic aSTR (autosomal short tandem repeats) read counts from NGS-DNA profiles, specifically for profiles generated from the ForenSeq™ DNA Signature Prep Kit, DNA Primer Mix B. Bright et al.'s model consists of three key components within the estimation of total allelic product-template, locus-specific amplification efficiencies, and degradation. In this work, we investigated the two mass parameters-template and locus-specific amplification efficiencies-and used MLE (maximum likelihood estimation) and MCMC (Markov chain Monte Carlo) methods to obtain point estimates to calculate the total allelic product. The expected read counts for alleles were then calculated after proportioning some of the expected stutter product from the total allelic product. Due to preferential amplicon selection introduced by the sample purification beads, degradation is difficult to model from the aSTR outputs alone. Improved modeling of the locus-specific amplification efficiencies may mask the effects of degradation. Whilst this model could be improved by introducing locus specific variances in addition to locus specific priors, our results demonstrate the suitability of adapting Bright et al.'s allele peak height model for NGS-DNA profiles. This model could be incorporated into continuous probabilistic interpretation approaches for mixed DNA profiles.

Human DNA extraction from highly degraded skeletal remains: How to find a suitable method?

Retrieving DNA from highly degraded human skeletal remains is still a challenge due to low concentration and fragmentation, which makes it difficult to extract and purify. Recent works showed that silica-based methods allow better DNA recovery and this fact may be attributed to the type of bones and the quality of the preserved tissue. However, more systematic studies are needed to evaluate the efficiency of the different silica-based extraction methods considering the type of bones. The main goal of the present study is to establish the best extraction method and the type of bone that can maximize the recovery of PCR-amplifiable DNA and the subsequent retrieval of mitochondrial and nuclear genetic information. Five individuals were selected from an archaeological site located in Catalonia-Spain dating from 5th to 11th centuries AD. For each individual, five samples from different skeletal regions were collected: petrous bone, pulp cavity and cementum of tooth, and rib and limb bones. Four extraction methods were tested, three silica-based (silica in-suspension, HE column and XS plasma column) and the classical method based on phenol-chloroform. Silica in-suspension method from petrous bone and pulp cavity showed the best results. However, the remains preservation will ultimately be the key to the molecular result success.

Reconstructing full and partial STR profiles from severely burned human remains using comparative ancient and forensic DNA extraction techniques

Thermal degeneration of the DNA molecule presents a special challenge to medico-legal investigations since low DNA yields, fragmented DNA molecules, and damaged nucleotide bases hinder accurate STR genotyping. As a consequence, fragments of severely burned human remains are often not amenable to standard DNA recovery. However, current ancient DNA (aDNA) extraction methods have proven highly effective at obtaining ultrashort DNA fragments (∼50 bp) from degraded palaeontological and archaeological specimens. In this study, we compare DNA yields and STR results obtained from two established aDNA and forensic DNA extraction protocols by sampling multiple skeletal elements recovered from victims (n = 23) involved in fire-related incidents. DNA yields and STR results suggest an inverse correlation between DNA yield and STR quality and increasing temperature. Despite the rapid thermal destruction of DNA at high temperatures, we generated higher quality full and partial STR profiles using the aDNA extraction protocol across all burn categories than the forensic total bone demineralization extraction method. Our analysis suggests adopting aDNA extraction methods as an alternative to current forensic practices to improve DNA yields from challenging human remains.

Statistical modeling of STR capillary electrophoresis signal

BACKGROUND

In order to isolate an individual's genotype from a sample of biological material, most laboratories use PCR and Capillary Electrophoresis (CE) to construct a genetic profile based on polymorphic loci known as Short Tandem Repeats (STRs). The resulting profile consists of CE signal which contains information about the length and number of STR units amplified. For samples collected from the environment, interpretation of the signal can be challenging given that information regarding the quality and quantity of the DNA is often limited. The signal can be further compounded by the presence of noise and PCR artifacts such as stutter which can mask or mimic biological alleles. Because manual interpretation methods cannot comprehensively account for such nuances, it would be valuable to develop a signal model that can effectively characterize the various components of STR signal independent of a priori knowledge of the quantity or quality of DNA.

RESULTS

First, we seek to mathematically characterize the quality of the profile by measuring changes in the signal with respect to amplicon size. Next, we examine the noise, allele, and stutter components of the signal and develop distinct models for each. Using cross-validation and model selection, we identify a model that can be effectively utilized for downstream interpretation. Finally, we show an implementation of the model in NOCIt, a software system that calculates the a posteriori probability distribution on the number of contributors.

CONCLUSION

The model was selected using a large, diverse set of DNA samples obtained from 144 different laboratory conditions; with DNA amounts ranging from a single copy of DNA to hundreds of copies, and the quality of the profiles ranging from pristine to highly degraded. Implemented in NOCIt, the model enables a probabilisitc approach to estimating the number of contributors to complex, environmental samples.

Systematic evaluation of STRmix™ performance on degraded DNA profile data

This study examined the DNA degradation modeling capacity of STRmix™, a widely implemented DNA interpretation software program. As a part of the CAL DOJ STRmix™ v2.4 validation, a large volume of STR profile data was generated from intact template DNA exposed to DNase I for a series of increasing time intervals. The resulting degraded profile data was analyzed with STRmix™ v2.4, and the efficacy of the analysis was assessed, both in terms of how the degradation modeling parameter values from the STRmix™ analysis compared to ground truth values, and how the weight-of-evidence statistics calculated for degraded profiles compared to those calculated for corresponding intact profiles. An additional set of differentially degraded mixture data was generated in silico to further challenge the STRmix™ degradation model, as well as to determine the extent to which end-user adjustment of the model's application can assist in resolving analysis problems that arise when high levels of degradation are observed in a profile. This work demonstrates that the degradation model in STRmix™ is capable of addressing a wide range of degraded STR profile data. The assessment expands the range of samples that have been rigorously examined using probabilistic genotyping approaches, as called for by forensic advisory bodies such as the United States President's Council of Advisors on Science and Technology.

Accurate assessment of the weight of evidence for DNA mixtures by integrating the likelihood ratio

Several methods exist for weight of evidence calculations on DNA mixtures. Especially if dropout is a possibility, it may be difficult to estimate mixture specific parameters needed for the evaluation. For semi-continuous models, the LR for a person to have contributed to a mixture depends on the specified number of contributors and the probability of dropout for each. We show here that, for the semi-continuous model that we consider, the weight of evidence can be accurately obtained by applying the standard statistical technique of integrating the likelihood ratio against the parameter likelihoods obtained from the mixture data. This method takes into account all likelihood ratios belonging to every choice of parameters, but LR's belonging to parameters that provide a better explanation to the mixture data put in more weight into the final result. We therefore avoid having to estimate the number of contributors or their probabilities of dropout, and let the whole evaluation depend on the mixture data and the allele frequencies, which is a practical advantage as well as a gain in objectivity. Using simulated mixtures, we compare the LR obtained in this way with the best informed LR, i.e., the LR using the parameters that were used to generate the data, and show that results obtained by integration of the LR approximate closely these ideal values. We investigate both contributors and non-contributors for mixtures with various numbers of contributors. For contributors we always obtain a result close to the best informed LR whereas non-contributors are excluded more strongly if a smaller dropout probability is imposed for them. The results therefore naturally lead us to reconsider what we mean by a contributor, or by the number of contributors.

Characterization of degradation and heterozygote balance by simulation of the forensic DNA analysis process

Simulation experiments were used to show the impact of varying extraction efficiency, aliquot proportion, and PCR efficiency on the heterozygote balance of a range of diploid and haploid cells. Reducing either parameters introduces variance. It is well-known that the variance in heterozygote balance increases as the amount of DNA is reduced. Surprisingly the distribution is in fact diamond shaped — the variance start to decrease at very low amounts of DNA. Simulations suggest that pristine diluted DNA is an acceptable approximation in validations to infer heterozygote balance. However, the difference in distribution of the variance between diploid and haploid cell types may, under some circumstances, need to be considered in statistical models. Finally, we exemplify how simulations can be used to predict the outcome of PCR for degraded samples. Visualizing the predicted DNA profile as an electropherogram can help to identify the best approach for sample processing.

Evaluation of forensic DNA mixture evidence: protocol for evaluation, interpretation, and statistical calculations using the combined probability of inclusion

Background

The evaluation and interpretation of forensic DNA mixture evidence faces greater interpretational challenges due to increasingly complex mixture evidence. Such challenges include: casework involving low quantity or degraded evidence leading to allele and locus dropout; allele sharing of contributors leading to allele stacking; and differentiation of PCR stutter artifacts from true alleles. There is variation in statistical approaches used to evaluate the strength of the evidence when inclusion of a specific known individual(s) is determined, and the approaches used must be supportable. There are concerns that methods utilized for interpretation of complex forensic DNA mixtures may not be implemented properly in some casework. Similar questions are being raised in a number of U.S. jurisdictions, leading to some confusion about mixture interpretation for current and previous casework.

Results

Key elements necessary for the interpretation and statistical evaluation of forensic DNA mixtures are described. Given the most common method for statistical evaluation of DNA mixtures in many parts of the world, including the USA, is the Combined Probability of Inclusion/Exclusion (CPI/CPE). Exposition and elucidation of this method and a protocol for use is the focus of this article. Formulae and other supporting materials are provided.

Conclusions

Guidance and details of a DNA mixture interpretation protocol is provided for application of the CPI/CPE method in the analysis of more complex forensic DNA mixtures. This description, in turn, should help reduce the variability of interpretation with application of this methodology and thereby improve the quality of DNA mixture interpretation throughout the forensic community.

Electronic supplementary material

The online version of this article (doi:10.1186/s12863-016-0429-7) contains supplementary material, which is available to authorized users.

Uncertainty in the number of contributors in the proposed new CODIS set

The probability that multiple contributors are detected within a forensic DNA profile improves as more highly polymorphic loci are analysed. The assignment of the correct number of contributors to a profile is important when interpreting the DNA profiles. In this work we investigate the probability of a mixed DNA profile appearing as having originated from a fewer number of contributors for the African American, Asian, Caucasian and Hispanic US populations. We investigate a range of locus configurations from the proposed new CODIS set. These theoretical calculations are based on allele frequencies only and ignore peak heights. We show that the probability of a higher order mixture (five or six contributors) appearing as having originated from one less individual is high. This probability decreases as the number of loci tested increases.

Sex determination in highly fragmented human DNA by high-resolution melting (HRM) analysis

Sex identification in ancient human remains is a common problem especially if the skeletons are sub-adult, incomplete or damaged. In this paper we propose a new method to identify sex, based on real-time PCR amplification of small fragments (61 and 64 bp) of the third exon within the amelogenin gene covering a 3-bp deletion on the AMELX-allele, followed by a High Resolution Melting analysis (HRM). HRM is based on the melting curves of amplified fragments. The amelogenin gene is located on both chromosomes X and Y, showing dimorphism in length. This molecular tool is rapid, sensitive and reduces the risk of contamination from exogenous genetic material when used for ancient DNA studies. The accuracy of the new method described here has been corroborated by using control samples of known sex and by contrasting our results with those obtained with other methods. Our method has proven to be useful even in heavily degraded samples, where other previously published methods failed. Stochastic problems such as the random allele drop-out phenomenon are expected to occur in a less severe form, due to the smaller fragment size to be amplified. Thus, their negative effect could be easier to overcome by a proper experimental design.

Evaluation of parameters in mixed male DNA profiles for the Identifiler® multiplex system

The analysis of complex DNA mixtures is challenging for forensic DNA testing. Accurate and sensitive methods for profiling these samples are urgently required. In this study, we developed 11 groups of mixed male DNA samples (n=297) with scientific validation of D‑value [>95% of D‑values ≤0.1 with average peak height (APH) of the active alleles ≤2,500 rfu]. A strong linear correlation was detected between the peak height (PH) and peak area (PA) in the curve fit using the least squares method (P<2e-16). The Kruskal-Wallis rank-sum test revealed significant differences in the heterozygote balance ratio (H(b)) at 16 short tandem repeat (STR) loci (P=0.0063) and 9 mixed gradients (P=0.02257). Locally weighted regression fitting of APH and Hb (inflection point at APH = 1,250 rfu) showed 92.74% of H(b) >0.6 with the APH ≥1,250. The variation of H(b) distribution in the different STR loci suggested the different forensic efficiencies of these loci. Allelic drop-out (ADO) correlated with the APH and mixed gradient. All ADOs had an APH of <1,000 rfu, and the number of ADO increased when the APH of mixed DNA profiles gradually decreased. These results strongly suggest that calibration parameters should be introduced to correct the deviation in the APH at each STR locus during the analysis of mixed DNA samples.

Analysis of allelic drop-out using the Identifiler(®) and PowerPlex(®) 16 forensic STR typing systems

Low-template (LT) DNA profiles continue to present interpretational challenges to the forensic community. Whether the LT contribution comprises the main profile, or whether it is present as the minor component of a mixture, ambiguity arises from the possibility that alleles present in the biological sample may not be detected in the resulting DNA profile. This phenomenon is known as allelic drop-out. This ambiguity complicates both the assessment of the potential number of contributors and estimation of the weight of the DNA evidence for or against specific propositions. One solution to estimating the weight of the evidence is to use a likelihood ratio (LR) that incorporates the probability of allelic drop-out P(DO) estimated for the specific evidence sample under consideration. However, although a vast repository of data exists, few empirical studies to determine allelic drop-out probabilities have been performed to date. Here we characterized patterns of allelic drop-out in single-source samples using both universal and run-specific analytical thresholds. Not surprisingly, we found fewer instances of apparent drop-out when using a lower (run-specific) detection threshold. Also, unsurprisingly, a positive correlation exists between allele drop-out and allele length, even in good quality samples. We used logistic regression to model the fraction of alleles that dropped out of a profile as a function of the average height of the detected peaks. The equation derived from the logistic regression model allowed us to estimate the expected drop-out probability for an evidentiary sample based on the average peak height of the profile. We show that the LRs calculated using the estimated drop-out probabilities were similar to those calculated using the benchmark drop-out probabilities, suggesting that the estimates of the drop-out probability are accurate and useful. This trend holds even when using the data from the PowerPlex(®) 16 typing system to estimate the drop-out probability for an Identifiler(®) profile, and vice versa. Thus we demonstrate that use of a LR that incorporates empirically estimated allelic drop-out probabilities provides a reliable means for extracting additional information from LT forensic DNA profiles.

Evaluation of mixed-source, low-template DNA profiles in forensic science

Enhancements in sensitivity now allow DNA profiles to be obtained from only tens of picograms of DNA, corresponding to a few cells, even for samples subject to degradation from environmental exposure. However, low-template DNA (LTDNA) profiles are subject to stochastic effects, such as "dropout" and "dropin" of alleles, and highly variable stutter peak heights. Although the sensitivity of the newly developed methods is highly appealing to crime investigators, courts are concerned about the reliability of the underlying science. High-profile cases relying on LTDNA evidence have collapsed amid controversy, including the case of Hoey in the United Kingdom and the case of Knox and Sollecito in Italy. I argue that rather than the reliability of the science, courts and commentators should focus on the validity of the statistical methods of evaluation of the evidence. Even noisy DNA evidence can be more powerful than many traditional types of evidence, and it can be helpful to a court as long as its strength is not overstated. There have been serious shortcomings in statistical methods for the evaluation of LTDNA profile evidence, however. Here, I propose a method that allows for multiple replicates with different rates of dropout, sporadic dropins, different amounts of DNA from different contributors, relatedness of suspected and alternate contributors, "uncertain" allele designations, and degradation. R code implementing the method is open source, facilitating wide scrutiny. I illustrate its good performance using real cases and simulated crime scene profiles.

Time-dependent changes in DNA stability in decomposing teeth over 18 months

OBJECTIVE

The objective was to use a dual quantitative and qualitative approach to analyze the dental DNA degradation produced by the passage of time since tooth death under controlled environmental conditions.

MATERIALS AND METHODS

Sixty human teeth were stored at room temperature for 0, 1, 3, 6, 12 or 18 months post-extraction. DNA quantification was determined by real-time quantitative PCR using a Quantifiler(TM) kit. DNA quality was assessed by the allelic dropout ratio between the smallest and largest loci obtained after STR genotyping and using an AmpFlSTR® Identifiler™ PCR kit. We also evaluated differences of DNA concentration related to gender and tooth position.

RESULTS

DNA concentration significantly reduced in 1 month post-extraction, stabilized between 1-12 months post-extraction, but decreased again at 18 months post-extraction. Interestingly, a significant reduction of the allelic dropout ratio (DNA quality) was only detected at 18 months post-extraction.

CONCLUSIONS

Stability of dental DNA decreased over time, differently affecting the amount and quality of the DNA in a time-dependent process over the first 18 months post-extraction. These results have a potential use in post-mortem intervals in human teeth in controlled environmental conditions.

Allelic drop-out probabilities estimated by logistic regression--further considerations and practical implementation

We discuss the model for estimating drop-out probabilities presented by Tvedebrink et al. [7] and the concerns, that have been raised. The criticism of the model has demonstrated that the model is not perfect. However, the model is very useful for advanced forensic genetic work, where allelic drop-out is occurring. With this discussion, we hope to improve the drop-out model, so that it can be used for practical forensic genetics and stimulate further discussions. We discuss how to estimate drop-out probabilities when using a varying number of PCR cycles and other experimental conditions.