Figure 1 . Flowchart of AIR algorithm
The AIR is an Artificial Intelligence-based protein structure Refinement method, which is constructed using a multi-objective particle swarm optimization (PSO) protocol.
We use multiple energy functions as multi-objectives so as to correct the potential inaccuracy from a single function. Given several initial models for one sequence, AIR takes each initial structure as the particle, with the process of structure refinement, the particles will also move around. The quality of current particles (structures) will be evaluated by three energy functions, and the non-dominated particles will be put into a set called Pareto set. After the iteration converges, the particles from the Pareto set will be screened and part of them will be outputted, which are the final refined structures.
Figure 2. Refinement case of AIR in the CASP 13 competition for target R0979. This target is the first oligomeric refinement target in CASP. On R0979, AIR's TOP1 output model achieves the best TM-align (0.80) and the best MAMMOTH (6.85) scores among all the 113 refined models submitted by 27 groups in the refinement category. The GDT_TS is improved by 2.99 compared to the initial model through AIR refinement process.
Di Wang, Ling Geng, Yu-Jun Zhao,Yang Yang, Yan Huang, Yang Zhang, and Hong-Bin Shen. Artificial intelligence-based multi-objective optimization protocol for protein structure refinement. Submitted. 2019