Identification and classification of proteins by FTIR microspectroscopy. A proof of concept

Abstract

FTIR spectroscopy is well known for its molecule fingerprinting capability but is also able to differentiate classes in complex biological systems. This includes strain typing and species level identification of bacterial, yeast or fungal cells, as well as distinguishing between cell layers in eukaryotic tissues. However, its use for the identification of macromolecules such as proteins remains underexplored and rarely used in practice. Here we demonstrate the efficacy of FTIR microspectroscopy coupled with machine learning methods for rapid and accurate identification of proteins in their dry state within minutes, from very small quantities of material, if they are obtained in a pure aqueous solution. FTIR microspectroscopy can provide additional information beside identification: it can detect small differences among different purification batches potentially originating from post-translational modifications or distinct folding states. Moreover, it distinguishes glycoproteins and evaluate glycosylation while detecting contaminants. This methodology presents itself as a valuable quality control tool in protein purification processes or any process requiring the utilization of precisely identified, pure proteins.