Ancient proteins preserved in archaeological and palaeoanthropological materials show great promise in studying the past. Proteins have been found to preserve in the archaeological record beyond the reach of ancient DNA [1,2], with applications ranging from studying diet and disease from dental calculus [[3], [4], [5]], to food items preserved in pot crusts [6], to reconstructing evolutionary relationships between extinct and extant species [[7], [8], [9]]. Palaeoproteomics also shows great potential in reconstructing the evolutionary history of our own species, as protein sequences preserved in enamel and other skeletal tissues have been shown to be phylogenetically informative for different taxa in the human family tree [2,10].

However, proteins degrade over time. Diagenetic processes cause proteins to hydrolyze into shorter fragments, and induce chemical modifications to amino acids [11]. Some proteins are completely lost from the original proteomes, and the proteins that partly survive are fragmented and damaged, and often only represented by a few peptides [12]. As a result of these processes, a significant proportion of the phylogenetic information contained in proteins may thereby be unrecoverable, and efforts need to be made to recover as much as possible of what remains.

The majority of peptides recovered from archaeological bone and dentine specimens stem from collagen, as collagens make up the vast majority of the protein content of these tissues [[12], [13], [14]]. Furthermore, collagen preserves well over time in comparison to other proteins initially present in bone and dentine [12]. Some studies have previously indicated that collagens, however, have comparatively low rates of amino acid polymorphism accumulation [12], potentially limiting the amount of phylogenetically relevant information. Recovering a larger fraction of non-collagenous proteins (NCPs) would therefore be highly beneficial for Pleistocene phylogenetic studies utilising ancient protein sequences as their source of information.

Previous research on both modern and Pleistocene skeletal elements has shown that using several proteases in parallel may significantly increase proteome size [[15], [16], [17], [18]]. Trypsin generates peptides with a positive C-terminal charge through efficient and highly specific cleavage C-terminal of lysine (K) and arginine (R), making it the preferred protease in mass spectrometry-based proteomics research [18]. Other proteases may not perform as efficiently, but when used in parallel, they may result in increased proteome sizes [15]. However, parallel digestion requires a larger amount of bone powder as starting material, which is then split between the different digestion conditions. As archaeological materials are a rare and limited resource, destructive sampling efforts should be kept to a minimum. In contrast, in modern and mediaeval contexts the sequential digestion of a proteomic extract with multiple proteases has been shown to result in larger proteomes [19, 20]. Thus far, however, only Lys-C, which is used to enhance trypsin cleavage, followed by trypsin has been widely adopted in the palaeoproteomics community [14]. Finally, in recent years new commercially available proteases have been developed and applied to modern and archaeological bone, such as LysargiNase [16,21] and ProAlanase [17].

Here, we explore the effect of protease choice on archaeological skeletal proteome size and protein sequence coverage, as well as the potential of sequential digestion with two proteases. In order to establish the efficiency of each protease in direct comparison, parallel single-protease digestion is first conducted on protein extract from three archaeological bone specimens, using trypsin, chymotrypsin, Glu-C, LysargiNase and ProAlanase. Thereafter, sequential digestion is conducted using the best performing proteases; chymotrypsin, Glu-C and ProAlanase, followed by trypsin.