Forensic Investigative Genetic Genealogy (FIGG) has emerged as a powerful tool for generating investigative leads in cases involving unidentified human remains and unknown perpetrators of violent crimes. The success of FIGG heavily relies on the generation of dense single nucleotide polymorphism (SNP) datasets suitable for genealogy database searching. While various genotyping strategies are currently in use, comparative evaluations based on authentic forensic case samples remain limited, with many studies relying instead on simulated datasets generated through downsampling of reads or generated from artificially treated DNA. In this study, we compared two approaches for SNP data generation: low-coverage whole genome sequencing (lcWGS) versus hybridization capture sequencing targeting approximately 1.3 million SNPs. The sample set comprised DNA extracts from authentic skeletal remains from routine cases and high-quality reference samples. WGS libraries were prepared and sequenced to a coverage of <0.001X to 1X, followed by genotype refinement and imputation using GLIMPSE2. In parallel, hybridization capture libraries were generated using the Twist aDNA SNP panel and analyzed using both threshold-based binary genotype calling and GLIMPSE2. All libraries were sequenced on a NovaSeq X platform. The resulting genotype datasets were evaluated in terms of SNP call rates and genotype concordance. Overall, the capture sequencing approach yielded higher SNP recovery when sufficient DNA input was available, whereas lcWGS, combined with genotype refinement and imputation, provided a robust alternative for samples with limited DNA input. For example, approximately one third of the included DNA extracts from authentic case samples produced SNP profiles exceeding 100,000 SNPs using capture sequencing, whereas approximately two thirds of the samples surpassed this threshold when applying the lcWGS approach. Notably, the application of imputation improved SNP recovery across both data types, although with varying effects depending on data quality and coverage. Practical considerations relevant for implementation, including laboratory and bioinformatic workflows, hands-on time and cost estimates, were also assessed and will be presented. In summary, both approaches enabled generation of SNP profiles suitable for FIGG from challenging skeletal remains, with performance influenced by DNA quantity and quality. The application of genotype refinement and imputation markedly improved SNP recovery and dataset completeness across both methods, demonstrating their role in maximizing the utility of low-coverage SNP data. These findings provide empirical guidance for method selection in FIGG oriented workflows for casework samples.

Authors

• Andreas Tillmar (Department of Forensic Genetics and Forensic Toxicology, National Board of Forensic Medicine, Sweden)
• Adam Staadig (Department of Forensic Genetics and Forensic Toxicology, National Board of Forensic Medicine, Sweden)
• Maja Krzewińska (Centre for Palaeogenetics, Stockholm University, Sweden)
• Maja Sidstedt (National Forensic Centre, Swedish Police Authority, Sweden)
• Siri Aili Fagerholm (National Forensic Centre, Swedish Police Authority, Sweden)
• Ricky Ansell (National Forensic Centre, Swedish Police Authority, Sweden)
• Anders Götherström (Centre for Palaeogenetics, Stockholm University, Afghanistan)
• Daniel Kling (Department of Forensic Genetics and Forensic Toxicology, National Board of Forensic Medicine, Sweden)