Materials and methods

The examinations and genetic analyses were performed within a broader governmental project dedicated to the identification of migrant victims, coordinated by the Office of the Commissioner for Missing Persons of the Italian Government and LABANOF (Laboratory of Forensic Anthropology and Odontology), University of Milan. The project involves national law enforcement authorities and multiple academic institutions. The University of Florence along with its forensic molecular anthropology lab -IRIS (Infrastructure for the Research and Identification of unknown Skeletons)- is a part of this group [19]. Out of the approximately 200 analyzed skeletal elements, 40 (20 auditory ossicles and 20 petrous bones) from as many individuals from the shipwreck that occurred in the Mediterranean Sea on April 18, 2015, were used for this study. All samples included in this study were procured from disarticulated skulls, which had not yet been associated with any individuals and had been submerged at sea for approximately one year following the shipwreck. Upon recovery, all disarticulated remains were preserved in cold storage at zero degrees Celsius, followed by a cleaning process through maceration. The petrous bones were extracted post comprehensive cleaning, while the auditory ossicles were collected as promptly as possible during or subsequent to the cleaning of the skulls. To ensure the accuracy and reliability of the molecular analysis, the recommended criteria for studies involving highly degraded DNA [20–22] were strictly adhered to, and the entire DNA analysis process was conducted in a dedicated laboratory where no other biological samples, such as blood, saliva, and semen, are processed. All testing was performed with the aim of maximizing the possibility of identification of these victims.

Auditory ossicles sample preparation

Auditory ossicles were obtained directly from the middle ear by personnel from the University of Milan using forceps. On the recommendation of personnel from the University of Florence, the anvil was preferred when available; otherwise, the hammer was chosen. Figure 1 shows a hammer and anvil from which DNA was extracted as an example.

Fig. 1: An example of an anvil on the left and a hammer on the right. The anvil was used for DNA typing

To remove potential contaminants, the outer layer of auditory ossicles was gently cleaned with a swab moistened with DNA-free water. After cleaning, each sample was irradiated with ultraviolet light for 10 min. in a Biolink DNA Crosslinker (Biometra). The auditory ossicles were then weighed using a precision balance (Radwag, model: ps 2100/c/2), and DNA was extracted from a range of 11–29 mg of bone.

Petrous bones sample preparation

Sampling of the petrous bone was conducted by personnel from the University of Milan using a Stryker saw. This process completely removed the temporal bone region, including the mastoid process, external auditory meatus, and petrous bone. At the IRIS lab, to eliminate potential contaminants, the outer layer of each petrous bone was mechanically removed using a low-speed rotary sanding drill (Dremel® 300 series). Subsequently, each sample was irradiated with UV light for 45 min in a Biolink DNA Crosslinker (Biometra). As proposed by Sirak et al. [23]., approximately 50 mg of bone powder was sampled by micro-drilling the petrous bones in order to keep the petrous bone as intact as possible. Figure 2 shows petrous bone microsampling as an example.

Fig. 2: The micro-hole from which the bone powder was collected from DNA typing is indicated by the red arrow

After micro-sampling, the petrous bones underwent axial computed tomography (CT) analysis to assess the effectiveness of micro-sampling techniques in reaching the cochlea, which is a crucial site for DNA analysis [14]. This evaluation was conducted at the Galeazzi Orthopedic Institute in Milan, employing the Newtom 7G. DICOM-formatted files were analyzed using Slicer 4.13 software [24].

DNA extraction and quantification

DNA was extracted from all samples following a published silica-based protocol [25] and eluted in 100 µl of TE buffer (10 mM Tris, 1 mM EDTA) with TET buffer (0.05% Tween-20). The extracts were then quantified using the Agilent 2100 Bioanalyzer System and the High Sensitivity DNA kit (Agilent Technologies). Although the TapeStation/Bioanalyzer platform is not human-specific and does not replace quantitative PCR methods routinely used in forensic casework, its use in this study was limited to assessing DNA fragmentation patterns and overall template availability prior to STR amplification. Therefore, quantification values were interpreted cautiously.

Short tandem repeat (STR) typing

STR typing of autosomal DNA was performed using the PowerPlex® Fusion 6 C System kit (Promega Corporation), capable of simultaneously amplifying 27 loci, meeting both CODIS and ESS recommendations. In addition to the samples, positive (2800 M Control DNA, Promega) and negative PCR and extraction controls were also amplified. Reaction was carried out on a SureCycler 8800 Thermal Cycler (Agilent Technologies) according to the manufacturer’s instructions. PCR products were separated via capillary electrophoresis on the SeqStudio Genetic Analyzer (Applied Biosystems) Cartridge v2, a 4-capillary electrophoresis system utilizing the universal POP-1 polymer. Genetic profiles were determined using the public domain, free, and open-source computer software Osiris, validated for clinical, forensic, and research applications [26–28]. The analytical threshold was set at 150 relative fluorescence units (RFU) for all auditory ossicles and for 9 out of 20 petrous bone micro-sampling. Profile completeness was defined based on successful amplification of all loci above the analytical threshold and consistency across replicate analyses, where performed, to support profile reliability. For the remaining 11 petrous bone samples, the threshold was set at 50 RFU, in accordance with forensic practice for highly degraded or low-template DNA samples. Lower thresholds may increase stochastic effects such as allelic dropout and drop-in in low-template DNA samples. However, replicate PCR amplifications were performed for selected samples, particularly for the 11 petrous bone samples requiring increased interpretative caution following the application of a reduced analytical threshold. Replicates were carried out from the same DNA extract, and consensus profiles were generated by considering only alleles consistently observed across independent amplifications, in accordance with established forensic DNA interpretation principles [29].

Statistical comparisons of RFU values per sample between auditory ossicles and petrous bone samples were performed using the non-parametric Mann–Whitney U test implemented in R. Effect size (r) was calculated to assess the magnitude of differences.