Distant kinship analysis has become increasingly important worldwide for the identification of historical human remains, including victims of past conflicts such as wars, as well as due to changing family structures and globalization. Traditional STR-based kinship analysis has limited power for resolving distant relationships. In contrast, chip-based approaches leveraging genomic segment sharing (e.g., shared cM, GD-ICS) from large SNP panels enable more refined inference of both the presence and degree of relatedness. However, these methods are constrained by their reliance on relatively high-quality and sufficient DNA, limiting their applicability to degraded or low-template samples such as historical human remains. To overcome these limitations, this study explores an NGS-based approach to extend genomic segment sharing analysis, building upon validated chip-based methodologies, and compares their performance.

A total of 35 pedigrees, comprising 41 kinship pairs, previously analyzed using a Korean SNP chip dataset, were re-analyzed with the ForenSeq Kintelligence kit. Among these, 39 pairs showed concordant kinship classifications, while two showed discrepancies: one pair was classified between the 4th and 5th degree rather than the expected 5th, and one pair did not yield a shared cM value. Unlike the chip-based approach, which distinguished parent-child and full sibling relationships, the NGS-based method did not differentiate these first-degree relationships and was limited to detecting relatedness up to the 4th degree. These differences are likely attributed to variations in SNP density, threshold criteria, algorithm sensitivity, and interpretation strategies.

Additionally, analysis of a Korean War casualty (~75 years postmortem) successfully confirmed a great-grandfather-great-grandchild relationship using GD-ICS inference, highlighting the practical applicability of this approach in challenging real-world forensic scenarios. This study highlights two key considerations in NGS-based kinship analysis: the required SNP density for reliable inference and the challenges associated with reduced SNP counts, providing empirical insight into these issues. Although NGS-based approaches are currently more effective for close relationships, their applicability to degraded and low-template samples was demonstrated in a real forensic context. When strategically combined with chip-based methods, these approaches offer a complementary framework that extends the utility of genomic segment sharing in forensic genetics.

Authors

• Yuna Oh (Division of DNA Analysis, Ministry of National Defense Agency for KIA Recovery & Identification, South Korea)
• Sohee Cho (Institute of Forensic Science, Seoul National University College of Medicine, South Korea)
• Soong Deok Lee (Institute of Forensic Science, Seoul National University College of Medicine, South Korea)