Negligible effect of host DNA on metagenomic analysis enables microbial ecology investigation in historical samples

Microbiome composition and function are shaped by its environment with major shifts induced by the intensified anthropogenic pressures over the past centuries. This time interval extends beyond experimental and longitudinal studies typically designed to investigate microbiome change. The extensive collection of historical samples available in museums and herbaria worldwide is an untapped resource for studying host-microbiome interactions over a prolonged time and in a global scale. Currently, their potential remains unfulfilled owing to incompatibilities with analytical pipelines and limited understanding of optimal classification parameters. Host DNA removal was deemed necessary for accurate taxonomic assignment of metagenomic reads, however, for many host species in historical samples it is impractical due to the absence of reference genomes. Here, we demonstrated that host DNA content has no significant impact on key microbial ecological metrics such as alpha- and beta-diversity. In addition, sequencing reads from historical samples are often highly fragmented due to degradation processes. Through k-mer analysis of genomics sequences from host species and their associated microbiomes, we observed that the shortest length of metagenomics read that can produce meaningful results for taxonomic profiling in historical samples is 21 bp. Overall, this study establishes a solid foundation for integrating extensive natural history collections into host- associated microbiome research, offering important insights into the anthropogenic impacts on microbiomes.