Figure 1

Overview of Multi-LZerD and this experiment. Multi-LZerD builds a protein complex by assembling pairwise docking predictions computed by LZerD. For each pair 54,000 models will be generated. The pairwise predictions can be clustered using a RMSD cutoff value of 5 or 10 Å or used for assembly without clustering. Pairwise docking models are assembled using GA. We tested different settings of the GA runs, with/without crossover operation, and different population sizes. After the final generation of GA, models can be clustered and refined, which we did not perform in this work.