New Ultra-Fast Computational Method Accurately Predicts Protein Spectra
Background
Accurately predicting the ultraviolet (UV) and fluorescence spectra of proteins is crucial for understanding their structure, function, and interactions. However, traditional computational methods are often too slow and resource-intensive for large biomolecules. This study addresses the critical need for an ultra-fast and accurate computational approach capable of handling large protein systems.
Results
The xTB-sTDDFT method demonstrated a remarkable reduction in computation time, cutting costs by more than 80% compared to previous methods. For the smaller polypeptides, oxytocin and GHRP-6, both UV and fluorescence spectra were successfully obtained, with errors between calculated and experimental values approximately 20 nm. This indicates a good level of agreement for these systems. For the much larger protein insulin, the method achieved even higher accuracy, with errors for both UV and fluorescence spectra consistently within 15 nm. Specifically, insulin's UV spectrum peak was calculated at λcal = 262 nm (experimental λexp = 277 nm, a difference of Δλ = 15 nm), and its fluorescence spectrum peak at λcal = 294 nm (experimental λexp = 304 nm, a difference of Δλ = 10 nm).
Why It Matters
This study establishes a robust and efficient theoretical model for ultra-fast calculation of electronic excitation spectra in large systems like proteins and other biomacromolecules. The combination of high accuracy and significantly reduced computational cost represents a major advancement in computational biology. This method could accelerate drug discovery and protein engineering by enabling rapid, reliable prediction of molecular properties, aiding in the design and characterization of novel peptide and protein therapeutics. Future applications could involve high-throughput screening of potential drug candidates or detailed analysis of complex protein interactions.