====
G-factor-based external bias limiter (GERBIL) is a scheme for regulating biased sampling
To regulate biased sampling a well-established structural indicator (the G-factor) was used to obtain structural information. Based on the G-factor, we proposed a scheme for regulating biased sampling, which is referred to as a G-factor-based external bias limiter (GERBIL). With GERBIL, the configurations were structurally validated by the G-factor during biased sampling. As an example of biased sampling, an accelerated MD simulation (aMD) was adopted in GERBIL (aMD-GERBIL), whereby the aMD was repeatedly performed by increasing the strength of the boost potential.
・NAMD 3.0 package
・Amber tools
・Python 3.x (3.7.8)
・MDAnalysis (2.1.0)
・matplotlib (3.3.1)
・pyEMMA (2.5.7)
・numpy (1.19.1)
・These scripts are written as premises for multi-node computers with the queuing systems (qsub).
0 Preparation
0-1 File preparation:put initial conformation(hoge.pdb) and parmeter file(hoge.prmtop) on ./gerbil_main/input/
0-2 Input parmeteres:write the number of residues and total number of atoms at the head of run_candi.sh
res_num=hoge name=hoge atoms=hoge
1 Conduct GERBIL
./run_gerbil.sh thresh_ini thresh_delta b_min b_max candi total_calculation_time(ns)
2 Calculated the FELs on the MSM constructions
2-0 Select initial structures for independently performed multiple long-time (5-ns) MD simulations from various configurations sampled by the aMD-GERBIL clustering_with_PCA.ipynb
2-1 Performed multiple long-time (5-ns) MD simulations run_cluster.sh $num_of_candidates
2-3 MSMs were constructed using the trajectories of the production MD runs the following scripts Get_cluster_PCA.ipynb cluster_MSM.ipynb
Launch run_gerbil.sh and load input files.
Takunori Yasuda, Rikuri Morita, Yasuteru Shigeta, Ryuhei Harada.
Center for Computational Sciences, University of Tsukuba
[email protected]