Towards personalised saliva spectral fingerprints: Comparison of mid infrared spectra of dried and whole saliva samples.

Affiliation

Ni D(1), Smyth HE(1), Gidley MJ(1), Cozzolino D(2).
Author information:
(1)Centre for Nutrition and Food Sciences, Queensland Alliance for Agriculture and Food Innovation, The University of Queensland, St Lucia 4072, Queensland, Australia.
(2)Centre for Nutrition and Food Sciences, Queensland Alliance for Agriculture and Food Innovation, The University of Queensland, St Lucia 4072, Queensland, Australia. Electronic address: [Email]

Abstract

The aims of this study were to compare two sample presentations (dry and whole) as well as the effects of both gender and age on the mid infrared (MIR) fingerprint spectra of human saliva. Unstimulated saliva was collected from 52 Female (31 subjects, aged 40.9 ± 14.6 year) and Male (21 subjects, aged 34 ± 11.8 year) participants, stored frozen, and subsequently thawed and analysed by MIR spectroscopy as whole and dried saliva, respectively. Data were analysed by means of principal components analysis (PCA) and partial least squares (PLS) to interpret and compare the effects of presentation (dry vs whole), age and gender on the MIR spectra of saliva. Interpretation of the MIR spectra of both whole and dried samples revealed specific characteristic and different spectral signals when gender and age were compared in the amide I and amide II of proteins (e.g. albumin) and DNA. While whole saliva analysis might be more convenient for rapid test, dried saliva spectra were more consistent across replicates, demonstrating greater ability to distinguish individual differences. The interpretation of the PCA and PLS loadings of both whole and dried saliva samples allowed identification of specific MIR regions associated with age and gender of participants between 1000 cm-1 and 1800 cm-1. In particular, the MIR regions associated with the absorption of polysaccharides, glycosylated proteins, and nucleic acid phosphate groups present in saliva were the most dominant. This paper demonstrates that MIR spectroscopy can be used to measure saliva samples and to interpret individual differences in participants due to age in either dry or whole samples. No clear trends were observed in the MIR spectra of the samples associated with gender when all samples were analysed together. However, PLS regression models were able to predict gender in a subset of samples having similar age. The approach described in this study shows promise for potentially using saliva as a tool in food studies (e.g. saliva interactions between food and consumers).