Fig. 4: Dynamical unstability of the P-TEFb/Tat/TAR-Delta system.
From: RNA regulation mechanisms study using physics-informed machine learning
a Schematic representation of the P-TEFb/Tat/TAR-Delta system. Blue and green denote Cdk9 and Cyclin T1, respectively, which together constitute the P-TEFb complex. Purple represents the Tat protein, while yellow represents the TAR-Delta RNA. The zoomed-in view highlights the Tat and TAR-Delta motif domains according to the secondary structures shown in (b). b Division of domains in the Tat protein and TAR-Delta RNA based on secondary structure. c Comparative analysis of backbone root-mean-square deviation (RMSD) between the P-TEFb/Tat/TAR (PTT) system (blue) and the P-TEFb/Tat/TAR-Delta (PTTD) system (green). The error bars represent the standard deviations of the backbone RMSDs across five independent trajectories of PTT and PTTD system. d Distribution of learned edges from neural relational inference learning between residues/nucleotides in the Tat/TAR-Delta complex MD simulations. e Distribution of learned edges among domains, obtained by aggregating the residues/nucleotides-level learned edges from the Tat/TAR-Delta complex MD simulations. The color bar represents the learned interaction strength. f Change in interaction strength between domains before and after bulge removal. The color bar shows this difference. g, h The interacting domains between Tat and TAR-Delta are identified and mapped from the learned interaction edges. Edge thickness indicates the interaction strength, corresponding to (e), while arrow direction indicates the directionality of a learned edge, representing the influence from the source domain to the target domain.
