Results and discussion

DNA was successfully extracted from all skeletal elements, and quality and quantity were initially assessed using the Agilent 2100 Bioanalyzer System. Although this instrument is not the standard for forensic quantification, its use provided valuable insights into DNA degradation patterns, informing subsequent STR analysis.

Auditory ossicles

Quantification values, assessed within the size range (70–500 bp) of amplified fragments, varied from 84.03 pg/µl to 876.32 pg/µl. Each sample produced a detectable signal within this range. In particular, the results highlighted that 15% of the samples showed quantification values below 200 pg/µl, with just one sample below 100 pg/µl. 40% of the samples had concentration values of approximately 200/400 pg/µl and the remaining 35% showed concentration values above 400 pg/µl with two samples (10% of the total) exceeding 690 pg/µl. Fig. S1 shows an example of quantification profiles of four auditory ossicles.

In addition to quantification information, electropherograms provided an assessment of the quality and size of the DNA extracts, showing a fairly homogeneous fragmentation profile over the range considered, with an increase in quantity around between 70 and 250 bp. Despite the DNA degradation and the presence of non-human DNA (Agilent analysis is not specific for quantifying human DNA), the results seemed to indicate that there was sufficient DNA to proceed with DNA typing. Therefore, STR typing was carried out on all extracts. According to quantification results, a complete and high-quality profile was obtained for all 20 analyzed auditory ossicles, with all 27 loci successfully typed. Figure 3 shows the reproduction of two STR profiles from auditory ossicles 14 and 15, for privacy reasons.

Fig. 3: The auditory ossicles reproduced STR profiles of samples 14 and 15, representing respectively the “best” and “worst” profile obtained among the four auditory ossicles quantification (Fig. S1)

As expected when amplifying degraded DNA samples, signal intensity decays as the size of the PCR product increases, and a “decay curve” is observed in which the peak height is inversely proportional to the amplicon length [30–33] (Fig. 4).

Fig. 4: Average RFU values calculated for each locus in the auditory ossicles’ profiles. Loci were arranged according to the fluorescent dye channels of the Powerplex® Fusion 6 C System kit. The decrease in RFU values with increasing amplicon size reflects the typical “decay curve” observed in degraded DNA samples, where shorter fragments amplify more efficiently than longer ones

As can be observed in Fig. 4, PCR products of approximately 100 bp (for example, D31358, D16S539, TH01, D8S1179, and DYS391) exhibited a higher signal compared to PCR products of approximately 400 bp (such as Penta E, Penta D, TPOX, D22S1045, and DYS570). This is typically because a greater number of DNA molecules remain intact within a smaller size range as opposed to a larger one. Nevertheless, even for the largest PCR products, notably high RFU values were obtained, with mean values spanning from 730 for the D22S1045 locus to 1781 RFU for TPOX. In particular, focusing on the evaluation of the largest individual PCR products (Table S1), it can be seen that these in all samples greatly exceed the analytical threshold of 150 RFUs. In addition, 55%, 25%, 60%, 30%, and 40% of the samples showed RFU values well above 1000 for Penta E, Penta D, TPOX, D22S1045, and DYS570, respectively.

These findings, supported by average RFU values ranging from 2585 to 14,255 RFUs, indicate that auditory ossicles maintain exceptional DNA preservation despite prolonged exposure to harsh environmental conditions, such as over one year submerged at sea. These values are significantly higher than those reported by Schwart et al. [12], who observed similar values only in cases without sign of putrefaction, whereas ossicles in our study performed well even under extreme degradation. This confirms that STR typing can be successfully applied to fully skeletonized remains recovered long after death, producing complete profiles of excellent quality. The robustness observed is likely linked to the extraction protocol optimized for highly degraded bone samples. Importantly, these results highlight that auditory ossicles are a largely underutilized resource for identification in humanitarian contexts. Their ability to yield complete, high-quality STR profiles while preserving cranial integrity supports their systematic inclusion in DVI workflows. By enabling genetic analysis without destructive sampling, ossicles safeguard anthropological information and align identification practice with ethical and cultural expectations, reinforcing dignity in mass fatality management [9, 10, 14].

Petrous bones micro-sampling

DNA concentrations, assessed within the size range (70–500 bp) of amplified fragments, ranged from 34.16 pg/µl to 779.01 pg/µl. Each sample produced a detectable signal within this range, although in some samples (samples 21 and 24 Fig. S2) the low amount of DNA made signal detection difficult. 45% of the samples (9 out of 20) registered concentrations over 300 pg/µl, including one sample with a quantification value significantly surpassing 700 pg/µl. Three samples presented quantification values under 100 pg/µl and another 40% had values ranging from 100 to 300 pg/µl.

Figure S2 shows an example of quantification of four micro-samples from petrous bones.

As can be observed in Fig. S2, a smear analysis was performed, and the estimated concentration values for the four samples were just over 30 pg/µl for sample 21, 150 pg/µl for sample 24, 300 pg/µl for sample 31, and approximately 200 pg/µl for sample 35. Additionally, the electropherograms were evaluated to assess the quality and size of the DNA extracts, revealing similar fragmentation profiles across the analyzed range. No increase in the quantification values was observed between 70 and 250 bp, with similar results along the entire range, in contrast to auditory ossicles, which showed a peak between 70 and 250 bp compared to 250–500 bp. A common feature observed in all petrous samples, regardless of DNA amount, was a distribution curve rising toward fragment lengths (sample 31 in Fig. S2) above 1000 bp, indicating possible exogenous DNA contamination.

Following quantification, STR typing was performed on all extracts. Complete STR profiles were successfully generated from 85% (17 out of 20) of the samples analyzed. To obtain reliable STR profiles, eleven of these samples required replication and consensus profiles were generated. Partial profiles with 18, 19, and 21 loci typed were obtained from the remaining 15%. Figure 5 shows examples of two STR profiles from petrous bones 24 and 35.

Fig. 5: The reproduced STR profiles from micro-sampling of petrous bone 35 and 24, representing respectively the “best” and “worst” complete profile obtained among the four micro-sampling quantification (Fig. S2)

In summary, all samples attained at least 18 STR loci, exceeding CODIS guidelines (13 loci) [34] and Italian standards (10 loci) [35].

Although high DNA quantification values were observed, this did not consistently yield complete, high-quality STR profiles. A limitation of the quantification approach used in this study is that the Agilent Bioanalyzer system does not distinguish between endogenous human DNA and exogenous DNA, including environmental or microbial contamination. This is particularly relevant for petrous bone samples, where electropherograms suggested the presence of high-molecular-weight DNA fragments, likely reflecting non-human DNA. As a result, quantification values may overestimate the amount of amplifiable human DNA, which could explain the discrepancy observed between DNA quantity and STR profile quality in some samples.

In some cases, replication was necessary, and the analytical threshold was set at 50 RFU, as for complex degraded traces. The use of a reduced analytical threshold increases sensitivity but may also increase stochastic effects; however, the use of replicate analyses and consensus profiling mitigates this risk and is consistent with forensic practice. Qualitative evaluation of electropherograms explains these results: graphs (Fig. S2) show low amounts of short DNA fragments, that is, those of the size expected for STR typing, but a notable presence of large fragments (evidenced by the hump present around 2000 bp), likely indicating contamination or non-human DNA. Signal strength decreased with increasing PCR product size, resulting in a more pronounced “decay curve” compared to auditory ossicles. This curve highlights that the peak height is inversely proportional to the length of the amplicon, as observed in Fig. 6.

Fig. 6: Average RFU values calculated for each locus (considering only detected alleles) in the petrous bone profiles. Loci were arranged according to the fluorescent dye channels of the Powerplex® Fusion 6 C System kit. A more pronounced decay curve is observed compared to auditory ossicles, indicating greater DNA degradation, with reduced amplification efficiency of longer STR loci

Average RFU values ranged from 268 to 10,858 per profile, confirming that petrous bone DNA is degraded but sufficient for complete or near-complete STR profiles. Anatomical variability also influenced results. Previous studies [13, 36] showed that the inner-ear portion (zone C) provides optimal endogenous DNA. Our micro-sampling strategy [23] preserved cranial integrity by CT scans confirmed that sampling never reached zone C, remaining about 6 mm away (Fig. 7). These observations highlight that petrous bone micro-sampling can preserve cranial integrity but requires careful planning and anatomical assessment. Pre-sampling CT imaging is recommended to optimize DNA yield while maintaining minimally invasive principles. Even without reaching the cochlear region, our results demonstrate that micro-sampling provides genetic profiles suitable for identification, supporting its use when less invasive alternatives are required.

Fig. 7: Computed tomography of petrous bone micro-sampling 24. The figure, with the sampling point circled in red, allows the distance (indicated by red line) between the sampling point and the cochlea to be assessed

However, despite this recommendation to optimize DNA yield, our analysis demonstrated that STR typing can still be successfully performed using micro-sampling without altering cranial integrity, supporting its application in humanitarian identification workflows.

Pros and cons in the choice of sampling

A key limitation of this study is the lack of paired sampling, as auditory ossicles and petrous bones were not collected from the same individuals. This introduces inter-individual variability and limits direct comparability between the two sampling strategies. Therefore, the observed differences in DNA yield and STR performance should be interpreted with caution and not as definitive evidence of the superiority of one method over the other. The primary aim of this study was to evaluate the feasibility of minimally invasive sampling approaches in a real DVI context, rather than to perform a direct comparative assessment under controlled conditions.

In this section, we summarize the main advantages and limitations of auditory ossicles and petrous bone micro-sampling for humanitarian identification in mass fatality contexts. This is not intended as a point-by-point comparison, which would require sampling both skeletal elements from the same individual.

Sample collection and preparation: Both approaches can be considered minimally invasive, although auditory ossicles involve virtually no structural alteration compared to petrous bone micro-sampling. However, their small size makes them prone to being overlooked or lost during recovery. Ossicles are also simpler to clean and do not require pulverization, eliminating a step in the laboratory workflow and reducing contamination risk. This translates into faster processing and improves operational efficiency.Quantification results: DNA quantity and quality varied widely, from < 100 pg/µl to > 800 pg/µl for auditory ossicles and up to approximately 700 pg/µl for petrous bone micro-sampling. Generally, ossicles tended to yield higher DNA amounts in this dataset in the 70–500 bp range, which are compatible with STR loci and natural degradation patterns. In contrast, petrous bone samples displayed a more uniform distribution across the analyzed range, without a clear increase in the short-fragment region, and often showed a hump around 2000 bp, likely due to exogenous or non-human DNA introduced during sampling. This explains why high quantification values in petrous bones did not always correspond to complete STR profiles.In general, the results obtained from quantification suggested a higher amount of DNA in the auditory ossicles compared to the micro-sampling of petrous bones. These results were associated with complete and generally higher quality STR profiles in the auditory ossicles within this dataset, although direct comparison is limited by the study design. Therefore, although quantification is non-specific for human DNA, it can still provide an indication of the quantity and preservation state for subsequent multiplex PCR analysis.STR typing: Both methods exhibited the expected decay curve for degraded DNA, with peak heights inversely related to amplicon size (Figs. 4 and 6). Comparing these curves (Figs. 4 and 6), auditory ossicles appeared to yield more DNA suitable for STR typing than petrous bones micro-sampling under the conditions of this study when the same PCR input was used. This was evident from consistently higher peak heights across all markers. Moreover, DNA from ossicles appeared less degraded, as indicated by lower proportional decreases within each fluorophore compared to petrous bone samples, which showed more extensive degradation. Consequently, complete STR profiles of excellent quality were obtained from all ossicles without PCR replication, whereas petrous bone micro-sampling produced complete profiles in 17 of 20 cases, with 11 requiring replication. A statistical comparison of RFU values per sample showed significantly higher values in auditory ossicles (median = 6144) compared to petrous bone samples (median = 1879; Mann–Whitney U test, W = 308, p = 0.0029), with a moderate effect size (r = 0.47). These results support the observed differences in amplification performance, although they should be interpreted with caution due to the lack of paired sampling. These profiles enabled identification of 19 individuals recovered from the Mediterranean Sea through standard DVI comparison procedures. Overall, these findings support a flexible, context-driven approach to skeletal sampling, prioritizing minimally invasive strategies to maximize identification success while preserving human remains in humanitarian operations.

Operational considerations for minimally invasive sampling in DVI contexts

The results of this study support the structured integration of minimally invasive skeletal sampling strategies into disaster victim identification (DVI) workflows involving highly degraded remains. Based on the comparative performance of auditory ossicles and petrous bone micro-sampling, several operational considerations emerge.

Preservation of anatomical integrity. Whenever feasible, sampling strategies should prioritize skeletal preservation, particularly of cranial structures that are critical for anthropological, odontological, and pathological assessment. Auditory ossicles can be removed with negligible structural alteration and without compromising subsequent forensic examination. Petrous bone micro-sampling, when performed using controlled micro-drilling techniques and anatomical guidance, can preserve cranial integrity while still providing sufficient DNA for STR typing.Balancing invasiveness and diagnostic yield. Sampling decisions should consider the expected DNA yield in relation to the degree of skeletal disruption. In the present study, auditory ossicles consistently yielded complete STR profiles without the need for replication, supporting their use as a first-line option when available. Petrous bone micro-sampling produced complete or near-complete profiles in the majority of cases and represents a valid alternative when ossicles are absent or inaccessible. The moderate effect size suggests that, while auditory ossicles generally provide higher RFU values, petrous bone samples may still yield comparable results in selected cases. This supports a flexible, context-dependent sampling strategy rather than a strict preference for a single method.Replication and analytical thresholds in degraded samples. In highly degraded contexts, sensitivity must be balanced with reliability. The requirement for allele replication and the adjustment of analytical thresholds in selected cases reflect established forensic practice for degraded skeletal material. Conservative interpretation criteria remain essential to prevent erroneous inclusions or exclusions in DVI scenarios.Integration within multidisciplinary identification procedures. Minimally invasive genomic sampling should complement, rather than replace, anthropological, odontological, and contextual analyses. The approaches evaluated here are compatible with established DVI protocols and may enhance identification efficiency while maintaining structural preservation of remains.

Together, these considerations provide a practical framework for implementing minimally invasive skeletal sampling in forensic casework and mass fatality investigations involving degraded maritime remains.