Premium
1 H NMR metabolite fingerprinting as a new tool for body fluid identification in forensic science
Author(s) -
Scano Paola,
Locci Emanuela,
Noto Antonio,
Navarra Gabriele,
Murgia Federica,
Lussu Milena,
Barberini Luigi,
Atzori Luigi,
De Giorgio Fabio,
Rosa Maria Francesca,
d'Aloja Ernesto
Publication year - 2013
Publication title -
magnetic resonance in chemistry
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.483
H-Index - 72
eISSN - 1097-458X
pISSN - 0749-1581
DOI - 10.1002/mrc.3966
Subject(s) - chemistry , principal component analysis , identification (biology) , set (abstract data type) , chemometrics , proton nmr , analytical chemistry (journal) , artificial intelligence , pattern recognition (psychology) , chromatography , biological system , computer science , stereochemistry , biology , programming language , botany
In this feasibility study, we propose, for the first time, 1 H NMR spectroscopy coupled with mathematical strategies as a valid tool for body fluid (BF) trace identification in forensic science. In order to assess the ability of this approach to identify traces composed either by a single or by two different BFs, samples of blood, urine, saliva, and semen were collected from different donors, and binary mixtures were prepared. 1 H NMR analyses were carried out for all samples. Spectral data of the whole set were firstly submitted to unsupervised principal component analysis (PCA); it showed that samples of the same BF cluster well on the basis of their characterizing molecular components and that mixtures exhibit intermediate characteristics among BF typologies. Furthermore, samples were divided into a training set and a test set. An average NMR spectral profile for each typology of BF was obtained from the training set and validated as representative of each BF class. Finally, a fitting procedure, based on a system of linear equations with the four obtained average spectral profiles, was applied to the test set and the mixture samples; it showed that BFs can be unambiguously identified, even as components of a mixture. The successful use of this mathematical procedure has the advantage, in forensics, of overcoming bias due to the analyst's personal judgment. We therefore propose this combined approach as a valid, fast, and non‐destructive tool for addressing the challenges in the identification of composite traces in forensics. Copyright © 2013 John Wiley & Sons, Ltd.