Maria A. Nieves‐Colón, Andrew T. Ozga, William J. Pestle, Andrea Cucina, Vera Tiesler, Travis W. Stanton, Anne C. Stone

The tropics harbor a large part of the world's biodiversity and have a long history of human habitation. However, paleogenomics research in these climates has been constrained so far by poor ancient DNA yields. Here we compare the performance of two DNA extraction methods on ancient samples of teeth and petrous portions excavated from tropical and semi-tropical sites in Tanzania, Mexico, and Puerto Rico (N = 12). All samples were extracted twice, built into double-stranded sequencing libraries, and shotgun sequenced on the Illumina HiSeq 2500. The first extraction protocol, Method D, was previously designed for recovery of ultrashort DNA fragments from skeletal remains. The second, Method H, modifies the first by adding an initial EDTA wash and an extended digestion and decalcification step. No significant difference was found in overall ancient DNA yields or post-mortem damage patterns recovered from samples extracted with either method, irrespective of tissue type. However, Method H samples had higher endogenous content and more mapped reads after quality-filtering, but also higher clonality. In contrast, samples extracted with Method D had shorter average DNA fragments. Both methods successfully recovered endogenous ancient DNA. But, since surviving DNA in ancient or historic remains from tropical contexts is extremely fragmented, our results suggest that Method D is the optimal choice for working with samples from warm and humid environments. Additional optimization of extraction conditions and further testing of Method H with different types of samples may allow for improvement of this protocol in the future.