Time–frequency approach in the cluster assignment of amino acids based on their NMR profiles

The “inversion recovery experiment” is used as a tool in nuclear magnetic resonance spectroscopy for characterization and identification of basic important amino acids. This methodology is based on determination of “longitudinal relaxation time ( T 1 )”of Carbon-13 of these molecular structures which relates to assignment of each carbon atom of amino acids. Not only the chemical shifts and frequencies of carbon atoms are different, but also the relaxation times of them in scale of seconds or less are different, without much overlapping. Due to larger shift effects in 13 C NMR spectra and larger paramagnetic origin for carbon-13 and other differences, it has made us to rely on 13 C nucleus as main clue in this work rather than 1 H nucleus. This procedure has helped us to identify the amino acids in terms of both “frequency” and “time” of relaxation for each carbon atom simultaneously. Applying “average linkage” as an agglomerative clustering method to the feature vectors extracted from NMR spectra of amino acids, a hierarchical clustering is provided. The obtained clusters reveal notable relationships between amino acids in a same cluster. After a time gap, the proposed clusters of amino acids which have similarities and differences with traditional grouping of amino acids provide a new perspective on amino acids characterization and related studies such as defining descriptors for proteins and peptides based on their sequence information.