Zonkey: A simple, accurate and sensitive pipeline to genetically identify equine F1-hybrids in archaeological assemblages

- We propose a method for identifying F1-hybrids in archaeological contexts using aDNA.
- We introduce Zonkey, open-source software implementing this methodology for equids.
- This method allows for direct taxonomic identification when morphology is ambiguous.
- Simulations demonstrate high sensitivity and specificity for small amounts of DNA.
- We test 18 samples and identify seven mules in Byzantine and Roman assemblages.

Horses, asses and zebras, can produce first-generation F1-hybrids, despite their striking karyotypic and phenotypic differences. Such F1-hybrids are mostly infertile, but often present characters of considerable interest to breeders. They were extremely valued in antiquity, and commonly represented in art and on coinage. However, hybrids appear relatively rarely in archaeological faunal assemblages, mostly because identification based on morphometric data alone is extremely difficult. Here, we developed a methodological framework that exploits high-throughput sequencing data retrieved from archaeological material to identify F1-equine hybrids. Our computational methodology is distributed in the open-source Zonkey pipeline, now part of PALEOMIX (https://github.com/MikkelSchubert/paleomix), together with full documentation and examples. Using both synthetic and real sequence datasets, from living and ancient F1-hybrids, we find that Zonkey shows high sensitivity and specificity, even with limited sequencing efforts. Zonkey is thus well suited to the identification of equine F1-hybrids in the archaeological record, even in cases where DNA preservation is limited. Zonkey can also help determine the sex of ancient animals, and allows species identification, which advantageously complements morphological data in cases where material is fragmentary and/or multiple candidate equine species coexisted in sympatry.