Museum genomics has transformed the field of collections-based research, opening up a range of new research directions for paleontological specimens as well as natural history specimens collected over the past few centuries. Recent work demonstrates that it is possible to characterize epigenetic markers such as DNA methylation in well preserved ancient tissues. This approach has not yet been tested in traditionally prepared natural history specimens such as dried bones and skins, the most common specimen types in vertebrate collections. In this study, we developed and tested methods to characterize cytosine methylation in dried skulls up to 76 years old. Using a combination of ddRAD and bisulphite treatment, we characterized patterns of cytosine methylation in two species of deer mouse (Peromyscus spp.) collected in the same region in Michigan in 1940, 2003, and 2013–2016. We successfully estimated methylation in specimens of all age groups, although older specimens yielded less data and showed greater interindividual variation in data yield than newer specimens. Global methylation estimates were reduced in the oldest specimens (76 years old) relative to the newest specimens (1–3 years old), which may reflect post-mortem hydrolytic deamination. Methylation was reduced in promoter regions relative to gene bodies and showed greater bimodality in autosomes relative to female X chromosomes, consistent with expectations for methylation in mammalian somatic cells. Our work demonstrates the utility of historic specimens for methylation analyses, as with genomic analyses; however, studies will need to accommodate the large variance in the quantity of data produced by older specimens.

These are double digest RADseq libraries that have undergone bisulfite treatment for methylation analysis. The DNA was extracted from dried bone tissue (microturbinates) sampled from Peromyscus maniculatus and Peromyscus leucopus specimens collected between 1940 and 2016. The libraries were digested with SphI and MluCI and individually tagged using a combinatorial barcoding system with barcodes on the forward adapter and indices on the reverse primer. All samples contain a spike-in of fully unmethylated lambda phage DNA - these reads can be used to calculate the sample-specific bisulfite conversion rate. The raw sequence data is uploaded here, as well as a document linking specimen ID to the barcode and index sequences. Detailed specimen information is available in the supplementary material of the associated manuscript.