Fig. 2: Molecular docking and RNA-folding prediction results obtained with Abacus.
From: A universal programmable Gaussian boson sampler for drug discovery
a, The docking pair of PARP-CQ (see text). Abacus is encoded with a 24-node BIG and finds the maximum weighted clique, using 347 sample data and 100-iteration local searches. b, Another 28-node BIG constructed by the complex of TACE-TS (see text), using 254 sample data and 10 iterations. In a and b the colored spheres denote the pharmacophore points we considered in the experiment (cyan: hydrogen-bond acceptor, HA, ha; blue: hydrogen-bond donor, HD, hd; yellow: negative charge, NC; orange: aromatic, AR, ar; and we use capital letters to represent pharmacophore points in the protein and lowercase letters to represent pharmacophore points in the ligand.) The sphere meshes in the ligands are other possible pharmacophore points but are not considered in our experiments. The shaded region in panels a and b denotes the maximum weighted cliques, as determined by the experimental results. All the cliques found from GBS experiments are shown in the pie charts; the maximum weighted cliques in both cases have a major proportion in the experimental results obtained from QIVS. More details are found in Supplementary Information Sections 5 and 6. c,d, The GBS-based RNA-folding prediction results for two RNA sequences (in c, accession no. AH003339; in d, accession no. AB041850). The max weighted cliques are highlighted in colored regions; they represent the most likely folding structures. The four colored circles (A, G, U, C) in the RNA structure represent the four different bases and the gray dashed lines represent the false negative base-pair matching. The colors of nodes within the clique, outlined by solid black lines, correspond to the predicted stems (base-pairs with matching shadow colors) in the RNA folding structure. For example, in c, the yellow node corresponds to the predicted stem with base-pairs 10–25, 11–24, 12–23 and 13–22. The cliques encircled by dashed pink (c) and yellow lines (d) represent other folding structures, though they are less accurate. More details are shown in Supplementary Information Section 7.
