Statistical fingerprinting

Dearman, R.J. et al., Cytokine fingerprinting of chemical allergens species comparisons and statistical analyses, Food and Chemical Toxicology, 40, 107, 2002. 

Computational methods have been applied to determine the connections in systems that are not well-defined by canonical pathways. This is either done by semi-automated and/or curated literature causal modeling [1] or by statistical methods based on large-scale data from expression or proteomic studies (a mostly theoretical approach is given by reference [2] and a more applied approach is in reference [3]). Many methods, including clustering, Bayesian analysis and principal component analysis have been used to find relationships and "fingerprints" in gene expression data [4]. 

The simplest approach to this problem is to search a database for an identical , i.e., similar within certain tolerances, spectrum. This was developed for Infrared spectra (a technique ideally suited to such a fingerprinting method). The method was enhanced to include a more sophisticated statistical approach when applied to NMR spectra.In NMR spectra, variation in peak position due to concentration and temperature effects is larger than the peak width, and a more sophisticated approach is mandatory. In either case, the method is clearly one which yields limited or even confusing information for novel compounds. 

Methodology In Figure 1 a) and b) the principles of catalytic profiling analysis are explained Catalytic profiling analysis includes a set of test reactions which are very sensitive with respect to catalyst properties and/or the recipes of preparation. From activity and selectivity values measured for a set of test reactions (Fig. la) corresponding performance profiles (Fig. lb)) can be derived which can be understood as catalytic fingerprints for individual catalysts. Thus, performance profiles allow a statistical analysis of similarities (Fig. lb). 

McHard et aK (152) using plasma spectroscopy investigated 32 elements in Florida and Brazilian frozen concentrated orange juice samples. Using probability statistics they suggested the relative occurrence of the minor elements Ba, B, Ga, Mn and Rb as ratios to zinc could act as "fingerprint" indicators of the geographic source of a sample. No relationship to adulteration was implied. 

The mass-spectrometric fingerprint —that is, the abundance of up to 900 single properties in the form of m/z with specific intensities—was shown to be most sensitive to detect, prove, and visualize even minor differences between samples, by the use of appropriate statistical procedures. This is independent on the specific sample properties (dissolved/solid, fractionated/nonfraction-ated) and was shown to disclose agronomic (fertilizer, manure, or crop-specific impacts on SOM quality), and ecological (parent material-metabolite, consumer-food, plant-soil) interrelationships. 

Although the characterization of coal macerals on the basis of their fluorescence spectra is a recent innovation, it has already proven to be an excellent fingerprinting tool for the various macerals. In some cases, it is even more sensitive than normal petrographic analysis. The initial results of fluorescence spectral studies show that the various fluorescent macerals in single coals can be statistically discriminated on the basis of their spectral parameters and that even varieties of a single maceral type can be distinguished. Although the spectra obtained at this time are rather broad and not suitable for chemical structure analysis, the potential for structural analysis exists and may be realized with improvements in instrumentation. 

The use of a direct combined (or polyphasic) approach can create highly specific soil fingerprints from normal constituents. This, in addition to the application of appropriate statistical analysis, would make soil analysis a more effective tool for routine forensic work, thus considerably extending its applicability. Indeed, combinations of different data each with its own discriminatory potential may result in probabilities of association or disassociation that even surpass those of techniques such as human DNA. Initial work using a canonical variate analysis has shown discrimination between soil types can be improved by including more analytical data. Figure 11.11 illustrates... 

Enumerating the product space of combinatorial libraries and calculating the descriptors can be expensive computationally. Recently, Downs and Barnard [62] have described an efficient method of generating fingerprint descriptors of the molecules in a combinatorial library without the need for enumeration of the products. This method is based on earlier technology for handling Markush structures. An alternative approach is to use random sampling techniques to derive a statistical model of the property under consideration [63],... 

---