Systematic Bias in TD-DFT Absorption Predictions for Nucleic Acid Fluorophores

Le résumé fourni par la source

Fluorescent nucleic-acid reporters are central to sequence-specific sensing and imaging, making reliable prediction of the absorption maximum () an important design objective. Due to the potential accelerated design of new reporters using computational methods, the present work benchmarks the performance of time-dependent density functional theory (TD-DFT) against experimental values for a database of 116 chromophores that span literature-validated fluorescent nucleobase analogues and other small molecule probes. Three methodological variables are systematically examined: 1) the relationship between linear-response TD-DFT and the Tamm–Dancoff approximation (TDA), 2) basis-set convergence from double- through quadruple- quality with and without diffuse augmentation for two different families, and 3) the performance of 63 density functionals distributed across Jacob’s ladder. values computed with TDA are found to differ from full TD-DFT results by relatively small, systematic shifts and exhibit a near-linear relationship to the full response treatment, enabling straightforward post hoc calibration while retaining substantial gains in computational efficiency and numerical robustness. Basis-set effects are dominated by the double- to triple- step, and diffuse augmentation is essential, particularly for probes with charge-transfer or diffuse character. aug-cc-pVDZ provides an effective basis set for screening, whereas aug-cc-pVTZ offers a reliable approach for data refinement. Across the methods tested, B2PLYP emerges as the most accurate functional, with mPW2PLYP close behind, and M06 stands out as the best global hybrid. Nevertheless, the optimal method is not universal, being dependent on probe properties. Although lower-cost hybrid functionals such as M06 and B3LYP remain viable for neutral probes and near-UV applications, long-wavelength, cationic/anionic, and cyanine-like (charge-transfer based) chromophores remain the most challenging to describe computationally and are most reliably treated by SOS/SCS double-hybrid functionals, especially the range-separated variants. Overall, the recommended computational workflow offers a productive route for pre-synthetic color tuning of nucleic-acid fluorophores.

Ce résumé expose les affirmations des auteurs. BNTIC ne l’interprète pas comme une validation indépendante des résultats.