The case is extraordinary on its own terms, but its scientific value lies in a rare natural experiment. Because both divers disappeared together and rested at nearly the same depth on the same seabed, they shared an identical postmortem submersion interval of roughly thirty years. Yet their bodies decomposed in strikingly different ways. One set of remains was largely reduced to a disorganized skeleton, while the other retained skin, hair, and recognizable muscle masses in the lower limbs. Two bodies, one environment, one timeline, two completely different taphonomic outcomes.

Both sets of remains underwent a full forensic workup. Whole-body postmortem computed tomography provided detailed internal imaging before any incision was made, followed by external examination, autopsy, toxicological analyses, and genetic identification. The imaging revealed advanced skeletonization predominantly affecting the upper body in both individuals, a pattern consistent with the well-documented sequence of marine decomposition, in which the head, neck, and thorax, rich in orifices and exposed to scavengers and currents, break down earliest. But the pelvic regions and lower limbs told a different story, preserving tissues compatible with adipocere, the waxy, soap-like substance that forms when body fat hydrolyzes under specific conditions.

Adipocere is one of the most fascinating phenomena in forensic taphonomy. It develops when anaerobic bacteria break down adipose tissue into saturated fatty acids, chiefly palmitic and stearic acid, which can inhibit further decomposition and essentially mummify soft tissue from within. Its formation typically requires a moist, oxygen-poor, alkaline environment, conditions that certain microenvironments within a submerged body can provide even on the open seafloor. In this case, computed tomography quantified the phenomenon precisely: the pelvic tissues compatible with adipocere showed attenuation values ranging from −100 to −40 Hounsfield units, a radiological signature that helps distinguish preserved fatty material from ordinary soft tissue or bone.

The critical difference between the two divers appears to have been their diving equipment. The first individual, designated X1, wore a two-piece semi-dry wetsuit, which allows water to circulate against the skin and offers limited protection from the surrounding environment. The second, X2, wore a full dry suit, which seals the body off almost completely. After thirty years on the seabed, X1’s residual tissues were largely disorganized, consisting of yellowish-to-brown aggregates, whereas X2 showed far more extensive preservation of skin, hair, and discernible muscle masses in the lower limbs. The dry suit, the authors suggest, may have created a sequestered microenvironment around the body, slowing bacterial activity and scavenger access in the enclosed regions, a hypothesis consistent with earlier experimental work showing that body coverings significantly alter adipocere formation in aquatic settings.

This finding matters because forensic scientists routinely use decomposition scoring systems to estimate the postmortem submersion interval, or PMSI. Several such aquatic decomposition scores have been developed and validated, including applications on bodies recovered from the Northern Adriatic Sea, and they assume a broadly predictable relationship between visible decomposition stage and time underwater. A thirty-year case in which one body looks far more decomposed than another, despite identical submersion times, exposes the limits of that assumption. The macroscopic appearance of a body recovered from water, the authors conclude, cannot by itself support a reliable PMSI estimate, especially after prolonged submersion and in the presence of adipocere.

The case also demonstrates the resilience of forensic genetics over decades in the marine environment. Identification of X1 was achieved from muscle tissue, while X2 was identified from femoral bone, a reminder that even when soft tissue fails, compact bone can preserve DNA for extraordinarily long periods. Previous reports have documented successful genetic identification of submerged remains after 26 years of submersion and full STR profiles from bones immersed for decades in freshwater lakes. Toxicological analyses in both Mediterranean cases came back negative, closing the loop on the medico-legal investigation that had waited three decades for its subjects.

Depth adds another layer of complexity. At more than 100 meters, the Mediterranean seafloor is cold, dark, and oxygen-poor, conditions that dramatically slow decomposition compared with shallow, warm, sunlit waters. Research on submerged carrion in the Strait of Georgia at 170 meters has shown that faunal scavenging at such depths differs profoundly from shallow-water patterns, with different scavenger communities and slower colonization. Comparative taphonomic studies have also distinguished between sequestered environments, such as enclosed bays or vehicle interiors underwater, and non-sequestered open water, where currents and scavengers accelerate disarticulation. The two divers, lying 13 meters apart on the same slope, may have experienced subtly different currents, sediment exposure, or scavenger activity, compounding the effect of their different suits.

Historical precedents hint at how long preservation can last underwater. German forensic literature from 1991 described body changes after fifty years in water, and case reports from Poland and elsewhere document submerged remains recovered 24 years after death. But the Nice case is unusual in offering a matched pair: two individuals from the same accident, same sea, same decade-long interval, examined with modern imaging and genetics. That pairing transforms an anecdote into evidence, allowing researchers to isolate variables such as clothing and microenvironment with a clarity that no laboratory experiment with animal models can fully replicate.

For forensic practitioners, the message is sobering and practical at once. When a body is recovered from water, the visible state of preservation is a starting point, not a conclusion. Radiological findings such as Hounsfield unit ranges for adipocere, contextual information about depth, temperature, clothing, and sequestration, and genetic identification from the most resilient tissues all need to be weighed together. Thirty years after two divers disappeared into the Mediterranean, their remains have finally given their families names, and given forensic science a benchmark that will shape how the next generation of underwater deaths is interpreted.

Subject of Research: Marine forensic taphonomy and postmortem submersion interval estimation in two divers recovered after 30 years

Article Title: Two divers submerged in the Mediterranean Sea for 30 years: persistence of adipocere and implications for postmortem submersion interval estimation

Article References: Bernardi, C., Amoretti, N., Isaac, A., Nogueira, L., Temma, J., Leccia, C., & Alunni, V. (2026). Two divers submerged in the Mediterranean Sea for 30 years: persistence of adipocere and implications for postmortem submersion interval estimation. International Journal of Legal Medicine. https://doi.org/10.1007/s00414-026-04038-y

Image Credits: AI Generated

DOI: 10.1007/s00414-026-04038-y

Keywords: adipocere, postmortem submersion interval, marine taphonomy, forensic genetics, postmortem computed tomography, forensic medicine, Mediterranean Sea, decomposition, dry suit, DNA identification, Hounsfield units, scuba diving fatality

Short tandem repeats, or STRs, are stretches of DNA in which a short sequence of two to six base pairs repeats over and over. The number of repeats at any given locus varies from person to person, which is precisely what makes STRs the backbone of modern forensic DNA profiling. Because these repeat counts are inherited, one copy from each parent, they allow investigators to compare a crime scene sample with a suspect’s profile, or to establish biological relationships in paternity cases. But the statistics underpinning those comparisons are only valid if laboratories know how common each repeat variant, or allele, is in the relevant population. A profile match that looks overwhelming in one population may carry different weight in another, which is why population-specific frequency databases are a cornerstone of ethical and accurate forensic practice.

The research team, led by Brajesh Kumar and Amitabh Biswas of Galgotias University’s School of Forensic Sciences, together with colleagues from the Central Forensic Science Laboratory in Chandigarh, Ingenomics India, and the Regional Forensic Science Laboratory in Jabalpur, used the Ingenomics AutoProfiler STR Kit to amplify all 24 markers simultaneously. This six-dye multiplex system, which was independently validated for forensic applications, covers the expanded set of loci now favored internationally, including the highly variable SE33 and Penta E markers alongside familiar names such as FGA, vWA, and TH01. Genotyping was carried out by capillary electrophoresis, and the resulting allele calls were subjected to statistical analysis using established population genetics software, including Arlequin, GenAlEx, and PowerStats.

The headline numbers are striking. Across the 24 loci, the probability of match, which expresses the chance that two randomly selected unrelated individuals share the same genotype, ranged from 0.012 to 0.186. Inverted into its more familiar form, the power of discrimination ranged from 0.814 to 0.988, meaning the best markers could distinguish between any two individuals in the population with near certainty. SE33 topped the list with a discrimination power of 0.988, followed closely by Penta E at 0.983 and FGA at 0.962. At the other end of the spectrum, CSF1PO at 0.814 and D2S441 at 0.844 were comparatively modest performers, a pattern consistent with their lower allelic diversity in populations worldwide.

Polymorphism information content, a measure of how informative a marker is for linkage and identity analysis, ranged from 0.601 to 0.933, with the majority of loci exceeding 0.7, a threshold conventionally regarded as indicating high polymorphism. Gene diversity values, which capture the expected heterozygosity at each locus, spanned 0.666 to 0.939, confirming that the Santhal sample carries substantial genetic variability. Observed and expected heterozygosities were calculated for every locus, and tests against Hardy-Weinberg equilibrium were performed to verify that allele frequencies behave as expected under random mating, a critical quality check before any forensic statistic can be considered reliable.

For paternity testing, the study’s parameters matter even more directly. The power of exclusion, the probability that a randomly selected man would be excluded as the biological father of a child, ranged from 0.428 to 0.913 across individual loci, with SE33 and Penta E showing the strongest exclusion capacity. The typical paternity index varied from 1.667 to 11.667, again led by the most polymorphic markers, including SE33, Penta E, and D12S391. When the loci were combined, the results became extraordinary: the combined power of exclusion reached 0.999999999991, the combined paternity index stood at 3.51 multiplied by ten to the tenth power, and the combined probability of match fell to 2.41 multiplied by ten to the power of negative 28. The combined power of discrimination, rounded, was effectively 1. In practical terms, the chance that two unrelated Santhal individuals would share an identical 24-locus profile is so vanishingly small that a matching profile constitutes evidence of identity at a level few other forms of forensic evidence can approach.

Why does population-specific data of this kind matter so much? Forensic statistics rest on the product rule: the rarity of a full profile is estimated by multiplying the frequencies of the individual alleles, an operation that assumes statistical independence within the population under consideration. If a population has a history of endogamy, consanguinity, or founder effects, as many isolated or semi-isolated communities do, allele frequencies can drift in ways that inflate or deflate match probabilities. The Santhal, an Austroasiatic-speaking tribal community with a distinct demographic history in eastern India, are precisely the kind of population for which generic national databases may not be adequate. By generating a dedicated allele frequency dataset, the study gives forensic practitioners in Bihar and neighboring states a defensible statistical foundation when the source of a sample may belong to this community.

The work also contributes to the broader scientific picture of Indian genetic diversity. Previous studies have documented STR polymorphism in Chhattisgarh, Himachal Pradesh, northwestern India, and among the Sindhi population, and genomic research on Austroasiatic speakers has highlighted the role of landscape barriers and sex-specific admixture in shaping their genetic structure. Each new population dataset refines the map of how genetic variation is distributed across the subcontinent, and the Santhal data will allow comparisons that illuminate both forensic and anthropological questions. The authors note that the 24 loci are highly informative not only for individual identification but also for kinship analysis, extending their utility to disaster victim identification, missing persons investigations, and complex relationship testing.

The study was approved by the Institutional Ethics Committee of Apex Hospital in Varanasi, and the authors report no conflicts of interest and no external funding. All participants were unrelated volunteers who consented to contribute samples, and the authors acknowledge their role in enabling the research. The sampling strategy, restricted to unrelated individuals, is essential for Hardy-Weinberg testing and for ensuring that the frequency estimates reflect the general population rather than clustered family lineages.

As DNA evidence continues to expand its role in the Indian judicial system, studies like this one quietly underpin every statistic that appears in a courtroom report. A match probability of ten to the negative 28 is only as credible as the allele frequency table behind it, and that table must reflect the population to which a suspect or victim belongs. With the Santhal community now represented in the forensic literature by a rigorously characterized 24-locus dataset, investigators and courts in Bihar gain a resource that makes DNA statistics for this population both more accurate and more defensible. It is a reminder that the power of forensic genetics depends not just on the chemistry of amplification and detection, but on the patient, community-by-community work of measuring human variation itself.

Subject of Research: Forensic genetic profiling of autosomal STR markers in the Santhal tribal population of Bihar, India

Article Title: Allele frequency distribution and forensic efficiency parameters of 24 autosomal STR markers in the santhal tribal community of Bihar, India

Article References: Kumar, B., Biswas, A., Dixit, S., Sharma, S., & Shrivastava, P. (2026). Allele frequency distribution and forensic efficiency parameters of 24 autosomal STR markers in the santhal tribal community of Bihar, India. The Science of Nature, 113(6), Article 125. https://doi.org/10.1007/s00114-026-02176-2

Image Credits: AI Generated

DOI: 10.1007/s00114-026-02176-2

Keywords: forensic genetics, STR markers, Santhal, allele frequency, paternity testing, population genetics, India, DNA profiling, Hardy-Weinberg equilibrium, power of discrimination, tribal population, polymorphism

The research, published in the International Journal of Legal Medicine, analyzed 8,237 unrelated Argentine individuals genotyped at 22 autosomal short tandem repeat loci, the repetitive DNA sequences that form the backbone of forensic identification worldwide. Rather than simply reporting allele frequencies for this enormous cohort, the team led by Antonella Belén Penacino and José Alonso Aguilar-Velázquez asked a more fundamental question: how does the size of the sample shape what the database appears to contain? To answer it, they generated 1,000 random resampling replicates for cohort sizes ranging from 500 individuals up to the full 8,237, effectively simulating thousands of alternative databases of different sizes drawn from the same population.

The results reveal a striking asymmetry between two properties of forensic databases that are often conflated. Across the 22 loci, the full dataset contained 344 distinct alleles, and nearly half of them, 169 alleles or 49.1 percent, occurred at frequencies below 1 percent. These rare alleles are the hidden tail of human genetic diversity, and it turns out they are extraordinarily expensive to capture. Overall allele recovery climbed quickly with sample size: a 500-person database captured 74.9 percent of the allelic diversity observed in the full cohort, a 3,000-person database reached 90.7 percent, and by 7,000 individuals the figure stood at 98.5 percent. Rare-allele recovery followed a much shallower trajectory, reaching only 81.0 percent at 3,000 individuals and still just 90.4 percent at 5,000.

This divergence matters because the two quantities serve different forensic purposes. The combined forensic parameters that courts care most about, the combined match probability and the combined power of exclusion, remained remarkably stable across all sampling scenarios in the study. The combined match probability showed minimal variation regardless of how many individuals were sampled, and the combined power of exclusion stayed consistently high throughout. In practical terms, even a 500-person database produced paternity and identification statistics that looked very similar to those derived from the full cohort of more than 8,000. For many routine applications, the headline numbers of forensic genetics appear almost indifferent to sample size.

Yet the stability of those combined parameters conceals what is happening at the level of individual alleles. When a DNA profile from a crime scene contains an allele that has never been observed in the reference database, laboratories must apply a minimum allele frequency, a floor value that prevents the match probability from being calculated as zero. The accuracy of that floor, and of the frequency estimates for genuinely rare alleles, depends directly on whether the database has sampled enough people to encounter them. A database that has recovered only 81 percent of the rare alleles present in its population is systematically underestimating the diversity that real casework will encounter.

The locus-specific analyses added another layer of nuance. Saturation dynamics, the point at which additional sampling stops revealing new alleles, were strongly associated with the proportion of rare alleles at each locus. Highly polymorphic markers such as FGA and D21S11, which harbor many low-frequency variants, required substantially larger sample sizes to achieve complete allele representation than less variable loci. This means that a single sample-size criterion applied uniformly across a multiplex panel is inherently misleading: some loci saturate quickly while others continue yielding new alleles thousands of samples later. The authors’ finding suggests that adequacy assessments should be conducted locus by locus, weighted by the intended application of the database.

The Argentine context makes the study particularly significant. Argentina’s population reflects complex admixture among Indigenous American, European, and other ancestral contributions, and earlier work by some of the same research community, including studies of urban Argentine populations and regional databases from Patagonia and the central provinces, has documented meaningful genetic structure across the country. Building a reference database at this scale, with informed consent from all participants, provides the forensic community with a resource whose allele frequency estimates carry far narrower confidence intervals than the small regional datasets that have historically been used. It also offers a benchmark against which the sufficiency of smaller national databases elsewhere can be judged.

The methodological approach draws on concepts borrowed from ecology, where rarefaction and species-accumulation curves have long been used to estimate how much of a community’s biodiversity a survey has captured. The study’s citation of foundational work on individual-based rarefaction by Robert Colwell and colleagues signals this intellectual lineage: alleles at a forensic locus are treated much like species in an ecosystem, and the resampling replicates function as accumulation curves revealing how discovery slows as sampling proceeds. The team’s analytical toolkit included standard population genetics software such as Arlequin and the STRAF online platform for forensic STR evaluation, alongside the R statistical environment for the resampling analyses.

The implications for forensic practice are direct. International guidelines, including the revised recommendations for publishing genetic population data issued by leading forensic geneticists in 2017, have grappled with how large a population sample must be, and recent work by other groups has begun questioning conventional sampling guidance for highly polymorphic STR loci. The Argentine study provides the strongest empirical answer yet: the answer depends on the question being asked. If the goal is stable combined forensic parameters for routine match probability and paternity calculations, moderate sample sizes perform adequately. If the goal is comprehensive representation of allelic diversity, particularly the rare alleles that populate nearly half of the allelic spectrum, then even several thousand individuals may not suffice, and databases aiming at full allele recovery should plan for substantially larger sampling efforts.

Perhaps the most enduring contribution of the study is conceptual. By demonstrating that allelic-diversity recovery and forensic-parameter stability are distinct properties that respond differently to sample size, the researchers have given the forensic community a framework for evaluating population databases according to their intended use rather than a one-size-fits-all threshold. As DNA phenotyping advances, as new multiplex kits expand the number of loci typed, and as courts increasingly scrutinize the statistical foundations of DNA evidence, that distinction will only grow in importance. A database that looks statistically adequate on paper may still be missing half the rare genetic variants its population carries, and knowing exactly which questions a database can and cannot answer is now an empirical matter that studies of this scale are finally equipped to resolve.

Subject of Research: Sample size effects on allele diversity and forensic parameters of autosomal STR loci in the Argentine population

Article Title: Sample size effects on allele diversity, rare-allele recovery, and forensic parameters of 22 autosomal STRs in a cohort of 8,237 Argentines

Article References: Penacino, A. B., Carvajal-Pérez, C. E., Rangel-Villalobos, H., Elsztein, L. D., Puentes, P. A., Zapata, F. A., Becerra-Loaiza, D. S., Moreno-Ortiz, J. M., Penacino, G. A., & Aguilar-Velázquez, J. A. (2026). Sample size effects on allele diversity, rare-allele recovery, and forensic parameters of 22 autosomal STRs in a cohort of 8,237 Argentines. International Journal of Legal Medicine. https://doi.org/10.1007/s00414-026-04032-4

Image Credits: AI Generated

DOI: 10.1007/s00414-026-04032-4

Keywords: forensic genetics, STR loci, allele frequencies, rare alleles, sample size, population database, combined match probability, power of exclusion, Argentine population, genetic diversity, DNA profiling, International Journal of Legal Medicine

The team followed the PRISMA 2020 reporting guidelines, the international gold standard for systematic reviews, to ensure their search and selection process was transparent and reproducible. They queried three major bibliographic databases, Scopus, PubMed, and Web of Science, for studies published through May 23, 2025. The initial sweep returned 301 records, which the authors then screened against strict inclusion criteria. In the end, 28 studies made the cut, all published between 2013 and 2025 and written in English. That twelve-year window captures nearly the entire modern history of forensic appearance prediction, from the earliest validated eye-color panels to today’s massively parallel sequencing assays that interrogate hundreds of markers in a single reaction.

From these 28 studies, the reviewers identified thirteen distinct categories of facial and externally visible characteristics that researchers have attempted to predict from DNA. Three traits stood out as having accumulated the strongest genetic evidence: overall facial morphology, the presence or absence of freckles, and male-pattern baldness. Each of these has been linked to a comparatively large number of single nucleotide polymorphisms, or SNPs, the single-letter variations in the genome that serve as the workhorses of genetic prediction. The breadth of the list is itself revealing, spanning pigmentation traits such as eye, hair, and skin color, structural features like ear shape, and age-related characteristics, reflecting both the ambition and the fragmentation of the field.

The review’s clearest finding concerns pigmentation. For common pigmentary traits, prediction accuracy has climbed above 89 to 90 percent in well-characterized populations, a level of performance that has already earned some assays a place in operational forensic work. Systems such as HIrisPlex and HIrisPlex-S, which predict hair, eye, and skin color from a compact set of SNPs, have undergone formal developmental validation, and commercial kits built on similar principles are now used in missing-persons investigations in several countries. The biology behind this success is relatively straightforward: a handful of genes, including HERC2, OCA2, and MC1R, exert large and well-mapped effects on melanin production, so a modest panel of markers captures most of the heritable signal.

Facial morphology is another story entirely. The review concludes that predicting complex facial configuration, including overall face shape and three-dimensional landmarks, remains a major unsolved challenge. Unlike eye color, the geometry of a human face is shaped by thousands of genetic variants, each contributing a vanishingly small effect, a genetic architecture scientists describe as highly polygenic. Studies using genome-wide association approaches in European and East Asian populations have identified dozens of loci associated with facial dimensions, from nasal bridge height to jaw width, but together these variants explain only a fraction of the observed variation. Environmental factors, aging, sex hormones, and developmental noise further blur the genotype-to-face mapping. Early efforts to build DNA-based facial composites, dating back more than a decade, produced recognizable statistical associations but never the photorealistic sketches that early media coverage promised.

Between these two extremes sit traits of intermediate complexity. Freckling, which reflects both melanocyte biology and sun exposure, has yielded predictive models built on dozens of SNPs, with a Polish study demonstrating that freckle presence can be estimated with useful accuracy. Male-pattern baldness, driven largely by androgen signaling pathways, has likewise been modeled from genetic data, though predictions are inherently sex-specific and age-dependent. Ear morphology has attracted growing interest, with a large multi-population genome-wide association study of 136 quantitative ear traits uncovering eight novel loci, yet translating such discoveries into forensic-grade prediction remains a work in progress.

Perhaps the review’s most consequential warning concerns population transferability. Predictive models trained on one ancestry group routinely lose accuracy when applied to another, because allele frequencies and linkage patterns differ across populations. The authors highlight genetically heterogeneous and admixed populations, particularly Latin American cohorts, as a critical gap. Latin America’s populations are mosaics of Indigenous American, European, and African ancestry, often varying dramatically from one region, or even one neighborhood, to the next. A prediction model calibrated on Europeans may systematically misestimate eye or skin color in such groups. The review argues that population-specific optimization and standardized frameworks are essential before forensic DNA phenotyping can be deployed equitably, and cites work on eye, hair, and skin color prediction in Latin Americans as an early step in that direction.

Technology is evolving on the laboratory side as well. The studies surveyed span a range of genotyping platforms, from targeted single-base extension assays and microarrays to massively parallel sequencing, which can process degraded or trace samples that defeat conventional methods. Researchers have applied these tools to challenging forensic material, including ancient and Second World War skeletal remains, single cells, and touch DNA, and newer platforms such as nanopore sequencing are being explored for rapid, field-deployable phenotyping. Both established commercial assays and laboratory-developed in-house panels were validated across the reviewed literature, indicating a maturing but still fragmented methodological landscape, one in which different laboratories use different marker sets, statistical models, and reporting conventions.

The implications reach beyond the laboratory. Forensic DNA phenotyping occupies contested ethical and legal territory, because it generates probabilistic descriptions of unknown individuals, potentially including sensitive characteristics, from biological material left involuntarily. Early commentators on the field flagged these concerns, and they have only intensified as the technology approaches operational use. The review’s authors frame their synthesis as a reality check: the science is genuinely powerful for pigmentation traits, genuinely immature for facial shape, and genuinely dependent on population context for everything in between. For investigators, the practical takeaway is that DNA-derived appearance predictions are best treated as investigative leads, narrowing suspect pools rather than identifying individuals, and always accompanied by calibrated uncertainty estimates.

What emerges from this systematic review is a field at an inflection point. The molecular foundations are solid, the genotyping technology is increasingly robust, and the first validated assays are already contributing to casework. But the dream of reconstructing a face from DNA remains distant, blocked by the sheer polygenic complexity of facial geometry and by the genetic diversity of real human populations. The authors’ call for standardized, population-aware frameworks amounts to a roadmap for the next decade of research. Until that roadmap is followed, the most reliable portrait a crime-scene sample can offer is still, in most cases, a person’s coloring rather than their countenance, a distinction that matters both for solving crimes and for protecting the innocent from premature suspicion.

Subject of Research: Forensic DNA phenotyping: predicting human facial features from molecular markers

Article Title: Evaluation of the use of molecular markers for the prediction of human facial features for forensic purposes: a systematic review

Article References: Valverde-Vilcherrez, A., Budowle, B., & Neyra-Rivera, C. (2026). Evaluation of the use of molecular markers for the prediction of human facial features for forensic purposes: a systematic review. International Journal of Legal Medicine. https://doi.org/10.1007/s00414-026-03989-6

Image Credits: AI Generated

DOI: 10.1007/s00414-026-03989-6

Keywords: forensic DNA phenotyping, facial features, SNPs, pigmentation prediction, systematic review, HIrisPlex, facial morphology, population genetics, forensic genetics, molecular markers, Latin American populations, International Journal of Legal Medicine

A team led by Anqi Chen, Qi Yang, Suhua Zhang, and Chengtao Li at the Institute of Forensic Science of Fudan University in Shanghai systematically compared the two dominant sequencing strategies for reading the full mitochondrial genome. On one side stood Illumina short-read sequencing, the workhorse of modern genomics, which chops DNA into fragments of a few hundred bases and reassembles them computationally. On the other stood nanopore long-read sequencing performed on a QitanTech platform, which threads much longer DNA molecules through protein pores and reads them in real time. The researchers applied both technologies to samples from 177 individuals spanning 39 maternal pedigrees, a design that allowed them to check whether each platform called the same variants in mothers, children, siblings, and cousins who should, by the rules of maternal inheritance, share essentially identical mitochondrial genomes.

The technical heart of the comparison lies in how each platform prepares DNA for sequencing. The short-read workflow relied on multiplex PCR, in which many primer pairs amplify overlapping fragments of the mitochondrial genome in a single reaction. This approach is efficient but notoriously uneven: some regions amplify enthusiastically while others lag behind, producing a patchwork of coverage depth across the genome. The long-read workflow took a fundamentally different route, using a single-amplicon enrichment strategy that amplifies the entire mitochondrial genome in one long piece. The payoff was dramatic. Long-read sequencing achieved a median coverage uniformity of 100 percent, effectively flattening the coverage variability that plagues multiplex PCR approaches and ensuring that every position of the mitochondrial genome is read a comparable number of times.

Coverage uniformity is not merely an aesthetic virtue. In forensic work, uneven coverage can mask genuine variants or inflate the apparent frequency of sequencing errors, complicating the interpretation of mixed or degraded samples. The long-read approach delivered a second, equally important advantage: the ability to see through one of mitochondrial DNA analysis’s most insidious traps. Scattered throughout the human nuclear genome are fragments of ancient mitochondrial DNA that migrated into our chromosomes over evolutionary time. These nuclear mitochondrial DNA segments, known as NUMTs, are molecular fossils that can be mistakenly amplified and sequenced alongside genuine mitochondrial DNA, contaminating forensic profiles with phantom variants. Short reads, typically a few hundred bases long, often cannot distinguish whether a sequence originated from a true mitochondrion or from a nuclear NUMT that happens to match the primer binding sites. Long reads, by contrast, can span the homologous regions entirely, revealing the surrounding nuclear context and allowing analysts to exclude these impostors with confidence.

The short-read data exposed the scale of this problem in vivid detail. The researchers observed pronounced off-target alignment in their Illumina datasets, with spurious reads piling up on particular human chromosomes. Chromosome 17 and chromosome 2 stood out as hotspots, accumulating misaligned sequences because of spurious sequence homology between NUMTs embedded there and the mitochondrial targets being amplified. In a forensic laboratory, such off-target material is more than a nuisance; it consumes sequencing capacity, complicates bioinformatic filtering, and in the worst case can introduce false variants into a profile that might later be used to include or exclude a suspect or identify human remains.

Yet the story is not a simple tale of long reads winning. When the researchers compared the consensus haplotypes produced by the two platforms, the sequences agreed: both technologies called the same set of mitochondrial variants for each individual. The critical divergence emerged in a more delicate measurement, the detection of heteroplasmy. Heteroplasmy refers to the coexistence of two or more mitochondrial DNA variants within a single person, a natural consequence of the way mitochondria are inherited through a genetic bottleneck and copied independently within cells. In forensic science, heteroplasmy can be a double-edged sword. On one hand, a shared rare heteroplasmic variant can provide powerful additional evidence linking relatives or identifying remains. On the other hand, distinguishing a genuine low-level heteroplasmic variant from a random sequencing error requires exquisite measurement fidelity, because true variants may be present at frequencies of just a few percent.

Here the short-read platform revealed its enduring strength. Illumina sequencing exhibited higher accuracy for detecting low-level heteroplasmy, thanks to the sheer depth and low intrinsic error rate of its short reads, which allow genuine minor variants to be separated from background noise with statistical confidence. The long-read workflow, despite its many advantages, carried a heavier background-noise burden. When the team validated their results against the family pedigrees, they found that the current long-read pipeline produced noise signals that were misclassified as heteroplasmy at a 10 percent allele frequency threshold. In other words, sequencing artifacts occurring at low frequencies could masquerade as genuine mixed mitochondrial populations, a potentially serious problem in casework where a false heteroplasmy call could undermine the interpretation of evidence.

Rather than declaring one technology the victor, the researchers proposed something more pragmatic and arguably more useful for the forensic community: a complementary workflow that assigns each platform the task it performs best. Under this integrated strategy, long-read sequencing would be deployed for structural characterization of the mitochondrial genome and for excluding NUMT interference, exploiting its superior coverage uniformity and its ability to span ambiguous regions. Short-read sequencing would then take over for high-fidelity quantification of heteroplasmy, where its precision at low variant frequencies is unmatched. The two technologies, rather than competing, would function as checks on each other, each compensating for the other’s weaknesses.

The implications for forensic casework are substantial. Laboratories handling complex or low-input samples, such as touch DNA from a weapon, hair shafts without roots, or skeletal remains recovered years after death, often face exactly the conditions where these platform trade-offs matter most. A degraded sample may yield too little DNA for confident short-read analysis of NUMT-contaminated regions, while the same sample might benefit enormously from long reads that clarify structural ambiguities. Conversely, a case hinging on a rare heteroplasmic variant shared between a missing person and a maternal relative demands the quantification accuracy that only deep short-read sequencing can provide. The Fudan team’s pedigree-based validation offers laboratories a data-driven rationale for choosing, or combining, their sequencing strategies rather than relying on convention or convenience.

The study also reflects a broader shift in forensic genetics, as nanopore sequencing matures from an experimental curiosity into a validated platform for legal applications. Chinese-made nanopore devices from QitanTech are emerging as alternatives to the long-dominant Oxford Nanopore systems, and their performance in this head-to-head comparison demonstrates that the technology has reached a level of maturity where systematic forensic validation, not just proof-of-concept demonstrations, is possible. With funding from China’s National Key Research and Development Program and the National Natural Science Foundation of China, and computations performed on Fudan University’s CFFF platform, the work signals that the infrastructure for rigorous platform benchmarking is now in place. For a field in which a single base call can help identify a fallen soldier or exonerate the innocent, knowing precisely when to trust long reads and when to demand the precision of short ones is not a technical footnote. It is the difference between evidence that stands up in court and evidence that collapses under scrutiny.

Subject of Research: Comparison of short-read and long-read sequencing platforms for forensic mitochondrial genome analysis

Article Title: Pedigree-Based comparison of short-read and long-read sequencing for forensic mitochondrial genome analysis

Article References: Chen, A., Yang, Q., Cao, Y., Liu, Y., Yang, F., Zhang, S., & Li, C. (2026). Pedigree-Based comparison of short-read and long-read sequencing for forensic mitochondrial genome analysis. International Journal of Legal Medicine. https://doi.org/10.1007/s00414-026-04031-5

Image Credits: AI Generated

DOI: 10.1007/s00414-026-04031-5

Keywords: mitochondrial DNA, forensic genetics, long-read sequencing, short-read sequencing, nanopore sequencing, heteroplasmy, NUMTs, DNA profiling, pedigree analysis, Illumina, QitanTech, human identification

The system, called Spitz, was developed by a team of forensic scientists at the Civil Police of the Brazilian Federal District in Brasília, working with real casework material rather than curated laboratory samples. At its core is a YOLOv10-L deep neural network, a member of the You Only Look Once family of object detectors that has become one of the most widely used architectures in real-time computer vision. Unlike slower two-stage detectors that first propose candidate regions and then classify them, YOLO models process an entire image in a single pass, predicting bounding boxes and class probabilities simultaneously. That speed matters in forensic settings, where a single microscope slide can contain thousands of fields of view and laboratories face persistent backlogs.

Training the model required a dataset grounded in the messy reality of forensic practice. The researchers assembled 399 microscopic image patches drawn from actual sexual assault cases, all stained with the Christmas Tree stain, the red-and-green histochemical preparation that is a staple of forensic sperm identification. This choice of training data is significant. Staining artifacts, epithelial cells, debris, and variable staining intensity make real casework slides far harder to interpret than idealized reference images, and models trained only on clean laboratory data often falter when confronted with authentic evidence. By anchoring the training set in genuine case material, the Brasília team aimed to build a detector that would hold up under the conditions examiners actually face.

The performance figures reported for the detection model are impressive, though they also reveal the inherent difficulty of the task. Spitz achieved a precision of 94 percent, meaning that when it flags an object as a spermatozoon, it is almost always correct. Recall, the proportion of true sperm cells it manages to find, came in at 72.3 percent, and the mean average precision at a 50 percent intersection-over-union threshold reached 0.842. In practical terms, the system errs on the side of caution: it rarely cries wolf, but it does miss some cells. For forensic work, that trade-off is arguably the right one, because the system is designed not to replace the examiner but to guide them, and a human expert remains in the loop for final verification.

What elevates Spitz beyond a standalone detector is the workflow built around it. The model was integrated into a web interface that supports batch inference across entire slides, reconstructs the spatial layout of the microscope slide, and records the coordinates of every detected sperm cell. An examiner can then navigate the microscope directly to flagged locations for confirmation, rather than sweeping the slide blindly. This coordinate-based verification bridges the gap between automated image analysis and the downstream steps of forensic genetics, where precise localization of sperm cells can feed into techniques such as laser capture microdissection and sperm cell sorting for DNA profiling. The system effectively turns the microscope from a search instrument into a confirmation instrument.

To test whether the technology actually improves forensic practice, the team ran a blind comparative study with three experienced forensic experts who re-examined 30 sexual assault samples using both conventional manual examination and the AI-assisted approach. For each sample, the examiners recorded spermatozoa counts, assigned a classification of Not present, Rare, or Occasional, and logged examination time. The design directly targeted one of the most stubborn problems in forensic biology: inter-examiner variability. Previous research has documented substantial differences in how analysts assess evidence suitability and interpret the same slides, and such variability can affect whether a sample proceeds to DNA extraction at all.

The statistical results favored the AI-assisted workflow across nearly every measure. AI-assisted detection found significantly more spermatozoa than manual examination, and the counts showed lower variability between examiners. Inter-rater agreement, quantified with Fleiss’ Kappa, jumped from 0.698 under manual examination to 0.858 with AI assistance, while raw agreement rose from 73.3 percent to 86.7 percent. The coefficient of variation across classification categories dropped from a mean of 93.5 percent to 80.1 percent, indicating that examiners converged far more closely on how to categorize samples when the algorithm had already flagged candidate cells. Wilcoxon signed-rank tests underpinned these comparisons, lending statistical weight to what the numbers show.

Examination time told a more nuanced story. AI-assisted examinations averaged 6.39 minutes with a standard deviation of 0.59 minutes, compared with 6.93 plus or minus 0.70 minutes for manual examination, a difference that did not reach statistical significance with a p-value of 0.2621. The modest time saving suggests that the primary benefit of Spitz is not speed but quality: more sensitive detection and more reproducible conclusions. The authors also highlight a benefit that is easy to overlook in the statistics: reduced examiner exposure to the intense illumination of the microscope, an occupational burden for analysts who spend their careers hunched over eyepieces.

The study arrives amid a broader wave of machine learning adoption in forensic science. Earlier efforts explored deep convolutional networks for sperm detection on microscope slides, YOLOv5-based sperm cell detection, and automated detection for laser capture microdissection, while commercial platforms have begun offering AI-assisted sperm finding. At the same time, critical reviews of machine learning in forensic DNA profiling have cautioned that validation standards must be rigorous, that datasets are often sensitive and hard to share, and that algorithms must earn the trust of courts and examiners alike. Spitz contributes to this conversation by reporting a validation study built on real casework, blind comparison, and established statistical measures rather than laboratory-only benchmarks.

There are, of course, limits to what the current study demonstrates. The annotated dataset and model weights cannot be released publicly because the underlying images come from sensitive forensic casework protected by institutional data policy, although the code for the Spitz interface may be shared on reasonable request. The recall of 72.3 percent means the algorithm alone is not exhaustive, which is precisely why the system is framed as an examiner’s assistant rather than an autonomous judge. Still, the direction of travel is clear. A model trained on 399 image patches from genuine cases, wrapped in a workflow that respects the expertise of human analysts, outperformed those same analysts on sensitivity and consistency in a blind trial. For a field where the stakes are measured in justice delivered or denied, that combination of technological ambition and procedural humility may prove to be the most important finding of all.

Subject of Research: Deep learning-based automated detection of spermatozoa in forensic microscopy for sexual assault investigations

Article Title: Spitz: a YOLO-based deep learning system for automated spermatozoa detection in forensic genetic analysis

Article References: Humanes, A. C., Calil, A. L. A., Wawruk, H. D., Carneiro, L. D., Silveira, K. B., Yen, W. C., de Andrade Gomes, J., & Meirelles, A. L. S. (2026). Spitz: a YOLO-based deep learning system for automated spermatozoa detection in forensic genetic analysis. International Journal of Legal Medicine. https://doi.org/10.1007/s00414-026-04013-7

Image Credits: AI Generated

DOI: 10.1007/s00414-026-04013-7

Keywords: forensic genetics, deep learning, YOLOv10, spermatozoa detection, sexual assault evidence, automated microscopy, inter-rater agreement, Christmas Tree stain, object detection, DNA profiling, forensic science, computer vision

The technology at the heart of the study is the Yfiler Plus PCR Amplification kit, a multiplex system that simultaneously amplifies 27 Y-chromosome short tandem repeats, or Y-STRs. These markers are repetitive stretches of DNA in which the number of repeating units varies from person to person. Because the Y chromosome is passed intact from father to son, a man’s Y-STR profile constitutes a haplotype, a signature of his paternal lineage. This makes Y-STR analysis uniquely powerful in sexual assault casework, where male DNA may be mixed with abundant female material, and in tracing male relatives, since paternal lineages share haplotypes. The trade-off is that the discriminating power of any Y-STR set depends entirely on knowing how frequently each haplotype occurs in the relevant population. A profile that is vanishingly rare in Europe may be commonplace in Borneo, and without local reference data, forensic statisticians cannot calculate meaningful match probabilities.

To supply that missing context, the researchers typed all 257 samples at the 27 loci and computed the standard forensic parameters for each of the four groups: haplotype diversity, discrimination capacity, and random match probability. These metrics collectively express how well the marker set distinguishes unrelated men within each community. The results, according to the authors, demonstrate the reliability of the tested Y-STR loci for forensic casework in these populations. In practical terms, when an investigator recovers a male DNA profile at a crime scene in Sabah or Sarawak, the new datasets allow an evidence weight to be assigned that reflects the actual genetic structure of local populations rather than a rough proxy drawn from unrelated groups thousands of kilometres away. The data also feed into international resources such as the Y Chromosome Haplotype Reference Database, which aggregates population haplotype frequencies worldwide for exactly this purpose.

Beyond the courtroom, the study offers a window into the deep history of Borneo’s peoples. The island occupies a pivotal position in the Austronesian expansion, one of the great demographic events of human prehistory, in which seafaring populations spread from Taiwan through Island Southeast Asia and out into the Pacific over several thousand years. Sabah and Sarawak are home to a striking diversity of indigenous Austronesian-speaking groups, each with distinct languages, customs and histories of migration and interaction. The Murut are traditionally an inland people of the interior highlands, the Bajau are famed as seafarers of the Sulu and Sulawesi seas, the Kadazan-Dusun form the largest indigenous group in Sabah, and the Melanau inhabit the coastal lowlands of central Sarawak. Whether their paternal lineages mirror these cultural and geographic distinctions was one of the questions the genetic data could address.

The answer, at first glance, is largely no. Using analysis of molecular variance, a standard technique that partitions genetic variation into within-population and between-population components, the team found that 89 percent of the Y-STR variation resided within populations, with only 11 percent separating the four groups from one another. In other words, two men from different communities in Borneo are, at the resolution of these markers, almost as similar to each other as two men from the same community. This pattern of high within-group diversity and low between-group differentiation is common in populations with shared ancestry, substantial gene flow, or both, and it underscores a recurring lesson of population genetics: cultural boundaries do not always map neatly onto genetic ones.

Principal coordinate analysis, a method that visualises genetic distances between populations in two dimensions, placed all four Bornean groups in a cluster together with other Austronesian-speaking populations of Southeast Asia, including Malays, Kedayan and Iban. Intriguingly, this Southeast Asian cluster sat slightly apart from indigenous Austronesian-speaking populations of Taiwan, such as the Paiwan and Puyuma, who are often considered descendants of the likely homeland of the Austronesian expansion. Had the Bornean groups descended directly and recently from Taiwanese ancestors, one might expect a closer genetic affinity. Instead, the Bornean and other Southeast Asian groups form their own cohesive cluster, hinting at the complex layering of migrations, admixtures and drift that has shaped the region’s paternal gene pool over millennia.

Yet the authors are careful to temper any grand historical conclusions. They suggest that the observed clustering patterns may stem from the limited resolution of the 27-locus Y-STR marker set rather than reflecting long-term demographic processes or the Austronesian expansion model inferred from linguistic and archaeological evidence. Y-STRs mutate relatively rapidly, but a panel of 27 markers can only capture so much of a lineage’s history, and haplotypes that appear similar may in fact be related only distantly, a phenomenon forensic scientists call homoplasy. Distinguishing genuine shared ancestry from marker-level coincidence requires either more markers, such as the rapidly mutating Y-STRs developed in recent years, or whole Y-chromosome sequencing, or the addition of genome-wide autosomal data that samples both maternal and paternal heritage.

This caution aligns with a broader shift in the field. Earlier generations of studies relied on small panels of 9, 12 or 17 Y-STR loci, and many of the published datasets for Malaysian populations, including work on Malays, Orang Asli groups, and the Iban, Bidayuh and Melanau of Sarawak, used these older kits. The 27-locus Yfiler Plus panel adds seven rapidly mutating markers that substantially increase the power to differentiate even close paternal relatives, a critical capability when a suspect’s brother or cousin may be the true source of crime scene DNA. The Bornean study thus brings the region’s forensic reference data up to the current international standard, and it complements recent genomic surveys that have revealed the native inhabitants of Peninsular Malaysia and North Borneo to carry a complex, layered population history that simple migration models fail to capture.

The authors are explicit about the limitations of the present dataset and the path forward. They call for larger sample sizes and for data from other genetically uncharacterised indigenous populations in Sabah and Sarawak to achieve accurate ancestry and forensic parameter estimations. Borneo hosts dozens of indigenous groups whose genetic profiles remain unsampled, and small sample sizes can bias frequency estimates, particularly for rare haplotypes whose true population frequencies are hard to pin down. Ethical engagement is equally important: the study obtained approval from the Human Ethical Committee of Universiti Sains Malaysia and written informed consent from every participant, following the tenets of the Declaration of Helsinki, a reminder that building forensic databases in indigenous communities must proceed with community trust and transparency.

For now, the study stands as both a practical tool and a scientific prompt. Practically, it gives forensic laboratories in Malaysia and the wider region the population-specific statistics they need to interpret Y-chromosome evidence from four of Borneo’s indigenous communities with confidence. Scientifically, it adds a piece to the puzzle of how the Austronesian-speaking world was peopled, while honestly acknowledging that the answer will require finer genetic resolution and broader sampling. As sequencing costs fall and interest in the genetic heritage of underrepresented populations grows, the men of Sabah and Sarawak whose DNA underpins this work may prove to be contributors not only to criminal justice in Malaysia but to a far richer understanding of humanity’s seafaring past.

Subject of Research: Forensic Y-chromosome STR population genetics of indigenous Austronesian-speaking groups in Sabah and Sarawak, Malaysia

Article Title: Population data of 27 Y-STR loci in indigenous populations from Sabah and Sarawak, Malaysia

Article References: Chan, X. Y., Kofi, A. E., Edinur, H. A., & Abd Rashid, N. H. (2026). Population data of 27 Y-STR loci in indigenous populations from Sabah and Sarawak, Malaysia. International Journal of Legal Medicine. https://doi.org/10.1007/s00414-026-04012-8

Image Credits: AI Generated

DOI: 10.1007/s00414-026-04012-8

Keywords: Y-STR, Yfiler Plus, forensic genetics, population genetics, Borneo, Sabah, Sarawak, indigenous populations, Austronesian, Murut, Bajau, Kadazan-Dusun

The core problem the researchers confronted is structural to DVI work itself. Identification by DNA typically relies on comparing post-mortem profiles recovered from remains against ante-mortem references: DNA from a toothbrush, a razor, a hairbrush, or a personal item reliably handled by the missing person. That approach collapses when a catastrophe kills whole families at once, because there is no surviving household from which to gather personal effects, and because direct relatives who would normally donate reference samples may themselves be among the victims. The October 7 massacre produced precisely this scenario across multiple communities, leaving investigators with fragmented, sometimes burned or commingled remains and a shortage of the reference material on which standard kinship matching depends. The solution proposed in the study was to mine medical archives, where every biopsy, Pap smear, surgical specimen, and diagnostic swab represents a documented, dated, and individually attributable sample of a specific patient’s DNA.

The researchers call this strategy a self-direct-reference antemortem DNA database, a term that captures its key advantage: the archived specimen is not a proxy handled by the victim, such as a toothbrush that could carry someone else’s DNA, but the victim’s own biological material collected under clinical conditions and linked to a verified identity through medical records. That distinction matters enormously in forensic practice, where misattribution of a reference sample can derail an entire identification. A paraffin block from a 2015 surgery or a cytology slide from a routine screening test carries an unbroken chain of custody through the pathology laboratory, and the DNA within it belongs, with near certainty, to the named patient. In effect, decades of routine diagnostic medicine had unknowingly been building a national ante-mortem DNA repository, waiting for a method to unlock it.

Unlocking it, however, is far from trivial, and this is where the study makes its most valuable technical contribution. Formalin fixation, the standard preservative for surgical pathology specimens, cross-links proteins and fragments DNA into short pieces, often just a few dozen to a few hundred base pairs long. Paraffin embedding adds further chemical and thermal stress over years of storage. Forensic short tandem repeat (STR) profiling, the workhorse of human identification, normally prefers intact DNA and amplifies loci that can span several hundred base pairs, so degraded archival DNA poses a real risk of allele dropout, locus amplification failure, and partial profiles. The team therefore designed a controlled optimization experiment, treating the archived specimens not as a last resort but as a material class with its own processing requirements, to be characterized systematically rather than improvised case by case.

The experimental design was straightforward but rigorous. The 151 specimens fell into three categories: 131 histological samples in the form of FFPE tissue blocks and cytology slides, 8 diagnostic swabs, and 12 blood or serum samples, all collected between 2003 and 2023. The investigators varied four parameters and measured their effect on profiling success: the tissue source of the specimen, the DNA extraction method applied, the incubation protocol used during extraction, and the age of the sample. Extracted DNA was amplified with the Investigator 24plex STR kit, and success was quantified as the number of informative autosomal STR loci recovered, with a maximum of 21 usable loci defining a fully informative profile. Descriptive statistics and non-parametric inferential tests were then applied to determine whether differences in outcomes across sample types and methods were statistically significant, which they were, confirming that processing choices genuinely matter rather than being matters of laboratory preference.

The headline result is a range of success rates spanning 67 to 100 percent of mean locus recovery, depending on tissue type and processing optimization. In practical terms, even the worst-performing specimen categories yielded profiles informative enough to contribute to identification, while the best categories performed as reliably as fresh reference material. Blood and serum samples, being less chemically assaulted than formalin-fixed tissue, sat at the favorable end of the range, as did certain tissue types within the FFPE group. The finding that sample age alone did not doom a specimen is particularly striking: blocks two decades old could still yield usable profiles when the right extraction chemistry and incubation conditions were applied. For DVI planners, this transforms the mental model of a pathology archive from a static library of diagnostic slides into a dynamic genetic resource whose yield can be predicted and maximized.

The optimization variables identified by the study carry direct operational significance. Extraction method selection determines how efficiently cross-linked, fragmented DNA is released from fixed tissue, and the literature on FFPE-derived DNA for next-generation sequencing has long shown that different extraction chemistries produce markedly different yields and fragment length distributions. The study extends that insight into the forensic STR context, where the goal is not sequencing read depth but the recovery of a sufficient number of complete, correctly sized STR amplicons. Incubation protocols, including the temperature and duration of the digestion step, similarly influence whether the fragile archival DNA is liberated intact or further sheared. By testing these variables across a real, diverse specimen collection rather than a small proof-of-concept set, the researchers produced guidance that a forensic laboratory can adopt directly, choosing the optimal workflow for each specimen type before committing irreplaceable archive material.

Beyond the immediate Israeli context, the study’s implications reach into global DVI preparedness. International guidelines, including INTERPOL’s DVI framework, emphasize ante-mortem data collection as the bottleneck of every mass fatality response, and historical operations from the Brussels Airport attack to the decades-long Korean War Identification Project illustrate how reference scarcity prolongs identifications and deepens families’ anguish. The Israeli experience adds a new layer to that body of knowledge: national health systems worldwide hold enormous, already-identified biological archives, and the study demonstrates that these archives can be converted into ante-mortem DNA references with predictable, optimizable success rates. The authors frame their findings as actionable guidance for DVI preparedness, and the phrase is apt, because the practical lesson is that countries need not wait for a catastrophe to begin cataloging, validating, and integrating their medical specimen repositories into forensic contingency planning.

There are also important scientific and ethical dimensions that the study implicitly raises. Forensic genetics in Israel operates within a legal and cultural landscape in which religious considerations, including Jewish law’s sensitivities around the treatment of human remains and autopsies, shape identification practice, making DNA-based methods that minimize invasive procedures especially valuable. The use of archived clinical specimens for identification purposes sits at the intersection of medical confidentiality, data protection, and forensic necessity, and the involvement of hospital pathology institutes alongside police forensic scientists in this study reflects the kind of institutional partnership such an approach requires. The authors report no competing interests and no external funding, and the work emerged from operational necessity rather than a pre-planned research program, which lends it the credibility of methods forged under real casework pressure.

What makes this research resonate beyond forensic circles is the way it reframes ordinary medical infrastructure as a form of civil resilience. Every pathology laboratory in the world quietly preserves a record of the people it has served, in the form of tissue blocks and slides that are retained for years or decades for clinical and legal reasons. The October 7 massacre forced Israeli scientists to recognize that this record could serve the dead and their surviving relatives when every other avenue of identification had been exhausted. The study’s systematic demonstration that 67 to 100 percent of informative STR loci can be recovered from such material, given the right extraction and amplification choices, converts a desperate improvisation into a reproducible protocol. For the families of the victims, the science delivered something no other method could: certainty about the fate of loved ones. For the global forensic community, it delivers a tested blueprint for the next inevitable mass disaster, one that begins not in the morgue but in the quiet shelves of the hospital archive.

Subject of Research: Use of archived medical specimens as ante-mortem DNA references for disaster victim identification

Article Title: Archived medical specimens as antemortem DNA references for disaster victim identification, optimization study following the October 7 massacre

Article References: Ido, A., Avrahami, K., Feinberg, T., Itzhaki-Alfia, A., & Herman, Y. (2026). Archived medical specimens as antemortem DNA references for disaster victim identification, optimization study following the October 7 massacre. International Journal of Legal Medicine. https://doi.org/10.1007/s00414-026-04028-0

Image Credits: AI Generated

DOI: 10.1007/s00414-026-04028-0

Keywords: disaster victim identification, ante-mortem DNA, FFPE specimens, STR profiling, forensic genetics, mass casualty, October 7 attack, pathology archives, DNA extraction, International Journal of Legal Medicine, Israel Police, DVI preparedness

Forensic kinship testing has long relied on short tandem repeats, the repetitive DNA sequences that underpin standard DNA profiling. STRs are excellent for identifying individuals and close relatives such as parents and children, but their statistical power fades rapidly for more distant relationships. As forensic geneticists increasingly turn to sequencing technologies and dense SNP data, a central question has emerged: how many SNPs are actually needed to resolve a given degree of relatedness, and can computational methods extract more information from a moderate number of markers than classical likelihood approaches alone? The new study addresses that question directly, testing a panel of nearly 2,000 SNPs that had previously shown strong forensic value in Chinese Han populations but had not yet been validated in other groups.

The researchers focused their kinship analysis on the Chinese Yunnan Zhuang group, a Tai-Kadai-speaking ethnic minority from Yunnan province in southwestern China. Their goal was to determine whether the moderate-density SNP panel could distinguish first-degree relatives, which include parent-child and full sibling pairs, second-degree relatives such as grandparents, grandchildren, half siblings, and avuncular pairs, third-degree relatives such as first cousins, and unrelated individuals. To do this, they computed two complementary measures for every pair of individuals: the logarithm of the likelihood ratio, a classical forensic statistic that compares the probability of the observed genotypes under a kinship hypothesis versus an unrelated hypothesis, and the cumulative identity-by-state score, which tallies how many alleles two people share at each marker simply by counting matching states.

The results were striking for close relatives. The density curves of the Log10 likelihood ratio and the cumulative identity-by-state values for first- and second-degree kinships were completely separated from those of unrelated individuals, meaning that with these markers, no close relative would be misclassified as unrelated or vice versa. Third-degree kinship proved harder, as expected, because first cousins share on average only about 12.5 percent of their DNA and the overlap with unrelated pairs becomes substantial. When the team set Log10 likelihood ratio thresholds at minus 4 and 4 to separate third-degree relatives from unrelated individuals, the system achieved a power of 86.40 percent with an error rate of zero. In practical terms, the panel can confidently exclude unrelated pairs and correctly flag a large majority of cousin-level relationships without producing false kinship calls, a critical property in forensic and disaster victim identification contexts where a false positive can have serious consequences.

Classical likelihood statistics, however, were only half of the story. The team also trained machine learning classifiers to sort pairs of individuals into four categories: first-, second-, third-degree relatives, and unrelated pairs. Using the Log10 likelihood ratio and cumulative identity-by-state values as input features, they compared several algorithms and found that the K-Nearest Neighbor model performed best, achieving an F1 score of 0.9781 across all four kinship classes. The F1 score, which balances precision and recall on a scale from zero to one, indicates that the model made very few false assignments across the full spectrum of relationship types. This result demonstrates that even a simple, interpretable machine learning method can capture patterns in SNP sharing statistics that fixed thresholds miss, effectively learning the boundary regions where likelihood ratios for different relationship classes overlap.

Beyond kinship, the study tackled a second forensic challenge: biogeographic ancestry inference at fine scale. The researchers integrated genetic data from 135 populations spanning nine geographic regions, drawn from public datasets, and conducted comprehensive population genetic analyses. The Yunnan Zhuang group showed the closest genetic affinity with geographically adjacent populations, particularly the Dai, Miao, and Tujia minorities of southern China and the Kinh population of Vietnam in Southeast Asia. This pattern is consistent with the known population history of the region, in which Tai-Kadai, Hmong-Mien, and related language groups share deep ancestral connections across southern China and mainland Southeast Asia.

Principal component analysis, phylogenetic tree construction, and ADMIXTURE-based ancestry modeling all revealed significant genetic structure among East Asian subpopulations, with especially clear separation between minorities from southern and northern China. Principal component analysis projects individuals into a low-dimensional space defined by the axes of greatest genetic variation, allowing clusters corresponding to population groups to emerge visually. Phylogenetic trees summarize the branching relationships among populations based on genetic distance, while ADMIXTURE models each individual’s genome as a mixture of ancestral components. The fact that all three approaches, applied to a panel of only about 2,000 SNPs, recovered this structure suggests the marker set captures enough ancestry-informative variation to serve as a practical tool for subpopulation discrimination, not just a research-grade dataset.

To push the discrimination further, the team applied multinomial LASSO regression, a feature-selection method that shrinks the coefficients of less informative markers to zero, effectively screening the full panel down to the most ancestry-informative subset. This process yielded 688 SNPs capable of distinguishing five East Asian subpopulations. Using these selected markers, the researchers built two classification models based on different strategies. A partial least squares-discriminant analysis model organized around a hierarchical classification approach, which first splits populations into broad groups and then subdivides them, and an Elastic Net model using a flat multi-classification approach, which assigns each sample directly to one of the five groups, both achieved overall accuracies above 0.9. The convergence of two methodologically different pipelines on similarly high accuracy strengthens the conclusion that the 688-SNP subset carries genuine, robust ancestry signal rather than artifacts of any single algorithm.

The implications for forensic practice are considerable. A moderate-density panel of this size can be genotyped efficiently with targeted sequencing or microarray platforms, at a fraction of the cost and data volume of genome-wide arrays, yet it delivers kinship resolution approaching that of much denser datasets for relationships out to second degree, and usable resolution for third degree when combined with machine learning classification. At the same time, the panel’s ability to distinguish southern from northern Chinese minorities and to place an unknown sample within East Asian substructure could help investigators narrow the geographic origin of unidentified remains or refine investigative leads. The study also adds to a growing literature on machine learning in forensic genetics, where algorithms are increasingly used to squeeze additional inferential power from standard marker sets rather than simply expanding the number of loci typed.

The authors note that the SNP panel had previously demonstrated strong forensic application value in Chinese Han populations, and the present validation in the Yunnan Zhuang group extends its evidence base to another major Chinese ethnic group. The work was funded by the Guangdong Provincial Science and Technology Program and the National Natural Science Foundation of China, and the underlying data are available from the corresponding author upon reasonable request, subject to privacy and ethical restrictions. As forensic laboratories worldwide weigh the transition from STR-centric workflows to SNP-based and sequencing-based platforms, studies like this one provide a practical benchmark: roughly 2,000 well-chosen SNPs, analyzed with a combination of classical likelihood statistics and accessible machine learning classifiers, appear sufficient to resolve close family relationships and to map fine-scale ancestry across one of the most genetically structured regions of the world.

Subject of Research: Forensic kinship identification and East Asian population discrimination using moderate-density SNPs and machine learning

Article Title: Moderate-density SNPs combined with machine learning method driven kinship identification and East Asian subpopulations discrimination analysis

Article References: Wu, X., Liu, Q., Luo, L., Lan, Q., & Zhu, B. (2026). Moderate-density SNPs combined with machine learning method driven kinship identification and East Asian subpopulations discrimination analysis. International Journal of Legal Medicine. https://doi.org/10.1007/s00414-026-04011-9

Image Credits: AI Generated

DOI: 10.1007/s00414-026-04011-9

Keywords: forensic genetics, SNP panel, kinship identification, machine learning, K-Nearest Neighbor, likelihood ratio, ancestry inference, East Asian populations, Yunnan Zhuang, population genetics, LASSO, International Journal of Legal Medicine

A team led by Juan A. Sanchis-Gimeno of the Universitat de València, together with colleagues in Spain, Chile and the United States, has published a proposed solution in the International Journal of Legal Medicine. Their Kinship-Centred Continuous Identification Framework is designed for historical and post-conflict mass graves where the pool of possible victims is open or uncertain. The study, published on 24 September 2026, is explicitly a framework development exercise: the authors stress that they have not invented new laboratory techniques, but have instead woven established evidence and operational standards into a time-ordered system with explicit decision points, feedback pathways and governance requirements.

The central premise of the framework is what the authors call the closing kinship window. Skeletal degradation and the progressive loss of highly informative relatives are coupled constraints. As time passes, DNA recoverable from bone may decline, but more critically the family members whose genetic profiles are most useful for identification, such as children and siblings of the missing, grow older and eventually die. Once those first-degree relatives are gone, identification must rely on more distant kin, whose genetic contribution to a match is weaker and whose genealogical connections are harder to document. Every year of delay narrows the range of relationships that can yield a statistically defensible identification.

To build the framework, the researchers conducted a purposive evidence map covering literature from database inception to 9 May 2026, drawing on fields that rarely sit at the same table: legal medicine, forensic genetics, humanitarian forensic action, archaeology, anthropology, disaster victim identification and transitional justice. This synthesis revealed recurrent operational bottlenecks, and the authors distilled their findings into seven design principles and ten implementation components. The recurring failure they identified is structural rather than technical. Family reference collection, post-mortem analysis, database matching and re-analysis are typically organised as separate projects, often by different institutions with different mandates and timelines. Information that should flow between these stages instead pools in silos.

The framework begins before any soil is moved. It requires mandate clarification, so that the legal authority and scope of an excavation are settled in advance, along with the construction of a provisional candidate-victim list. Genealogical triage follows, prioritising which families should be approached first for reference samples based on how informative their kinship links are likely to be. This front-loading of genealogical work is a deliberate inversion of common practice, in which family sampling often starts only after remains are already in the laboratory, wasting precious time while the kinship window narrows.

Once excavation is under way, the framework incorporates commingling-aware sampling, a critical consideration in mass graves where bodies were often dumped together and skeletal elements from different individuals may be intermixed. Each sample is classified for profile informativity, and marker selection is question-led rather than routine. Instead of applying a single standard genetic test to everything, the framework directs analysts to choose the genetic markers, whether autosomal short tandem repeats, Y-chromosome markers, mitochondrial DNA or dense single nucleotide polymorphism panels suited to extended kinship analysis, that best answer the specific identification question posed by each sample and its candidate relatives.

Matching is then organised along two axes: direct matching against reference profiles from personal items or medical samples, and programme-wide kinship matching across the entire database of victims and relatives. Crucially, the framework insists on trained human interpretation of statistical results rather than blind reliance on software output, followed by multidisciplinary reconciliation in which genetic evidence is weighed alongside archaeological, anthropological and documentary findings before an identification is confirmed. The system also requires periodic re-examination of direct-reference options and of profiles that remain unresolved, so that new family samples or improved technologies can be brought to bear on cold cases within the same programme.

The case-generating example for the framework is Paterna Cemetery in Valencia, Spain, which contains mass graves associated with executions during the Francoist repression following the Spanish Civil War. Published programme-level evidence from Paterna illustrates why grave assignment, candidate lists and genealogies must remain revisable throughout an identification effort. Earlier meta-research by the same group on 15 mass graves at Paterna, covering 933 individuals, documented identification success rates, and a 2026 aggregate analysis of official exhumation reports revealed discrepancies between the individuals expected in each grave and those actually recovered. In other words, even official records about who lies where can be wrong, and a rigid identification pipeline built on fixed assumptions will propagate those errors.

The technical underpinnings the framework draws upon are well established in the literature. Studies from the Balkans demonstrated highly effective DNA extraction methods for skeletal remains and documented how typing success varies between skeletal elements, with the petrous portion of the temporal bone emerging as an exceptionally rich source of DNA. Guidelines from the International Society for Forensic Genetics govern the use of Y-chromosome, X-chromosome and mitochondrial markers in kinship analysis, and the validation of biostatistical software. More recently, extended kinship analysis using SNP capture and sequencing kits designed for investigative genetic genealogy has expanded the range of relationships that can be resolved, potentially softening the blow of the closing window, though the authors note that such approaches carry their own cost, throughput and governance considerations.

The authors’ ultimate recommendation reaches beyond methodology into institutional design. Legal medicine services, they argue, should move from episodic exhumation support toward accountable, consent-based and continuously updated identification infrastructure. Identification of the missing should not be a series of discrete projects that end when funding does, but a standing capability that maintains databases, revisits unresolved profiles and keeps genealogies current across decades. For the families of the missing, who live with what researchers describe as ambiguous loss, the difference between an episodic programme and a continuous one is not administrative detail. It is whether the remains of a parent or a child are ever named at all, and whether that answer arrives while a sibling or a daughter is still alive to receive it.

Subject of Research: A kinship-centred continuous identification framework for genetic identification of remains in historical and post-conflict mass graves

Article Title: The closing kinship window: a continuous identification framework centred on kinship for historical and postconflict mass graves

Article References: Sanchis-Gimeno, J. A., Schwab, M. E., Valenzuela-Fuenzalida, J. J., & Granite, G. (2026). The closing kinship window: a continuous identification framework centred on kinship for historical and postconflict mass graves. International Journal of Legal Medicine. https://doi.org/10.1007/s00414-026-04023-5

Image Credits: AI Generated

DOI: 10.1007/s00414-026-04023-5

Keywords: forensic genetics, mass graves, kinship analysis, DNA identification, humanitarian forensics, legal medicine, Spanish Civil War, Paterna Cemetery, family reference samples, disaster victim identification, transitional justice, skeletal DNA