Optimised strategies for read mapping and snp filtering of highly fragmented DNA from keratinised tissues

Keratinised tissues offer a non-invasive and readily accessible DNA source and are increasingly used in museum genomics, wildlife genetics, and conservation studies. However, DNA derived from these materials typically exhibits intermediate characteristics between high-quality genomic DNA and ancient DNA, including substantial fragmentation but comparatively low levels of DNA damage. Despite the growing importance of keratinised materials in forensics, existing workflows are not optimised for analysing DNA sequence reads from these materials. Here, we systematically evaluated and optimised key parameters in workflows for read mapping and variant calling to account for the distinctive properties of DNA obtained from keratinised tissues. By fine-tuning the bwa-aln alignment parameters (-l and -o), we substantially improved computational efficiency while maintaining mapping accuracy and reliability. In parallel, we refined single-nucleotide polymorphism (SNP) filtering criteria based on the empirical distribution and quality characteristics of keratinised-tissue-derived genomic DNA (kmDNA), adjusting the mapping quality threshold from > 40 to > 36. This modification maximised the retention of genuine genetic variants and significantly increased the number of usable SNPs. The resulting optimisation framework shows strong stability and reproducibility across diverse sample types and data qualities. It provides a methodological basis for applying kmDNA in population genetic analyses and large-scale genomic studies and highlights the need for data-specific optimisation of analytical parameters in other highly fragmented DNA datasets.