Accuracy and computational feasibility of transport calculations in molecular and biomolecular junctions

Charge transport calculations are central to the interpretation and design of molecular and biomolecular junctions, but their predictive power is limited by both the accuracy of the underlying electronic structure and the computational cost of treating open systems. Here, we assess the accuracy and computational feasibility of transport calculations across junctions of increasing complexity, from small rigid molecular bridges to extended biomolecular systems. We compare NEGF-based approaches combined with density functional theory (DFT) and higher-level GW calculations, with particular focus on electronic level alignment and on approximate treatments that enable simulations of larger junctions. For small molecular junctions, we show that post-SCF DFT+Σ corrections substantially improve level alignment and yield densities of states and transmission functions in good agreement with GW benchmarks at a much lower computational cost. For larger systems, including porphyrin and protein junctions, we evaluate projection-operator diabatization (POD)-based transport schemes as computationally efficient alternatives to full NEGF calculations. The all-to-all POD approach reproduces the main transmission trends qualitatively, whereas the Breit-Wigner approximation is reliable mainly for smaller systems with well-separated electronic states. Overall, our results define the regimes in which GW/NEGF, DFT+Σ/NEGF, and POD-based approaches are quantitatively reliable, qualitatively useful, or computationally impractical, and provide practical guidance for transport studies spanning molecular to biomolecular junctions.