Biomolecules include both small molecules such as metabolites, vitamins, and hormones and large macromolecules such as amino acids and proteins. In the case of proteins of high molecular weight, their three-dimensional arrangements enable them to fold into one or more specific spatial conformations driven by noncovalent interactions. Raman spectroscopy gives information about the selectivity, the sensitivity and the conformation of molecules. In single-molecule techniques such as fluorescence microscopy and atomic force microscopy allow for the direct observation of the dynamics of proteins. This avoids the complexities of averaging over heterogeneous dynamics in bulk biochemical assays. The most sophisticated and effective methods for classifying surface-enhanced Raman scattering spectra are based on predetermined library of spectra. They are called supervised methods and include linear discriminant analysis and partial least squares discriminant analysis. However, building a valid classification model requires the acquisition of a training set of samples.