Expedient Bayesian prediction of subfossil bone protein content using portable ATR-FTIR data

Abstract

Rapid and minimally destructive methods for estimating the endogenous organic content of subfossil bone save time, lab consumables, and valuable ancient materials. Fourier transform infrared (FTIR) spectroscopy is an established method to estimate bone protein content, and portable spectroscopes enable field applications. We review the ability of benchtop and portable FTIR indices to predict %N and %collagen from 137 bone specimens drawn from eight taxa. We also explore associations of these indices with the endogenous DNA content estimated for 105 specimens. Bulk bone elemental abundance and crystallinity index data reflect diagenetic alteration of these specimens, which come from a variety of depositional environments in four countries (Madagascar, Greece, Monaco, and Germany). Infrared (IR) indices from benchtop and portable units perform similarly well in predicting observed sample N content and collagen yields. Samples that include little collagen (0–5 wt%) tend to have similar IR index values, and we present a Bayesian approach for the prediction of collagen yields. Bone type best explains variation in target species DNA content (endogenous DNA being particularly abundant in petrosals), but low IR index values were consistently associated with minimal DNA content. We conclude that, although portable FTIR fails to distinguish collagen preservation among poorly preserved samples, a simple approach with minimal sample preparation can effectively screen bone from a variety of taxa, elements, and environments for the extraction of organics.