计算化学公社

标题: 求助：gromacs中FEP自由能扰动法计算溶剂化自由能如何调整λ使得出结果标准误差减小 [打印本页]

作者
Author: sxmn 时间: 2026-3-4 10:45
标题: 求助：gromacs中FEP自由能扰动法计算溶剂化自由能如何调整λ使得出结果标准误差减小
10个受体Y6，2000个溶剂分子。先能量最小化，按照dy-npt.mdp先平衡30n是，再按照 λ=0.00 0.05 0.10 0.15 0.20 0.25 0.30 0.35 0.40 0.45 0.50 0.55 0.60 0.65 0.70 0.75 0.80 0.85 0.90 0.95 1.00平衡2ns，命令行gmx bar -f md-*.xvg -b 1900 -o bar_results.xvg -oi bar_int.xvg -oh bar_hist.xvg输出结果
Detailed results in kT (see help for explanation):

lam_A  lam_B    DG +/-    s_A +/-    s_B +/- stdev +/-
   0    1 -272.07  0.61 11.34  0.28  187.41  1.48 25.09  1.16
   1    2 -440.54  1.18 17.80  0.81  690.85  1.70 50.02  1.94
   2    3 -906.71  1.54  -53.48  0.97 -290.93  1.36 156014528.45 7647697.46
   3    4 -495.94  0.49 12.73  0.54  296.24  1.27 27.17  1.73
   4    5 -496.72  1.13 11.11  1.36 11.25  0.87 7.97  1.68
   5    6 -108.31  1.06 19.04  1.98 53.83  1.43 50.51  1.67
   6    7  236.13  0.32  140.20  1.35  137.24  0.57 466964257839805244251308032.00 1516076403662126709601206272.00
   7    8  514.94  2.52  -31.62  2.15 -200.73  2.95 60.70  1.48
   8    9 1181.31  1.95  354.26  3.35  342.78  2.80 5716792558421376573856712818153600887592179769037529302712495112192.00 52788563421184399531582118199565905663974988866142271021076473136021504.00
   9    10  812.91  2.70  758.08  5.30  725.72  2.87 39639538688520239283873617945214995102474991474502967683328327205282704281504393837519704494845341457868712931203065999390428826582414108930396464873472.00  -nan
10    11  116.92  2.00  133.31  6.05  113.97  3.61 149814116281954959360.00 50891591938264730173440.00
11    12 35.67  1.86 70.85  1.92 58.52  1.83 38897252.10 2306728074.01
12    13 13.81  1.27 46.30  2.90 50.55  1.80 2512.41 434456.38
13    14 0.89  1.74 33.08  0.99 45.81  2.25 79.36  1.93
14    15  -11.24  2.20 31.11  1.53 38.14  1.96 65.42  2.25
15    16  -16.71  1.46 27.86  1.17 34.99  1.43 40.51  2.51
16    17  -20.54  0.99 29.96  1.73 31.72  1.13 64.05  3.09
17    18  -14.89  1.45 29.77  1.54 34.46  2.04 54.57  4.09
18    19 -7.52  2.44 41.95  1.64 39.41  2.44 80.52  3.60
19    20 18.31  2.69 79.33  3.74 68.48  2.77  981.21 86.30

WARNING: Some of these results violate the Second Law of Thermodynamics:
      This is can be the result of severe undersampling, or (more likely)
      there is something wrong with the simulations.

Final results in kJ/mol:

point    0 -    1, DG -678.64 +/-  1.52
point    1 -    2, DG -1098.84 +/-  2.94
point    2 -    3, DG -2261.65 +/-  3.83
point    3 -    4, DG -1237.03 +/-  1.22
point    4 -    5, DG -1238.99 +/-  2.83
point    5 -    6, DG -270.17 +/-  2.64
point    6 -    7, DG 589.00 +/-  0.79
point    7 -    8, DG 1284.44 +/-  6.27
point    8 -    9, DG 2946.58 +/-  4.86
point    9 -    10, DG 2027.67 +/-  6.72
point    10 -    11, DG 291.65 +/-  4.99
point    11 -    12, DG 88.97 +/-  4.65
point    12 -    13, DG 34.45 +/-  3.17
point    13 -    14, DG  2.21 +/-  4.35
point    14 -    15, DG -28.03 +/-  5.48
point    15 -    16, DG -41.69 +/-  3.63
point    16 -    17, DG -51.22 +/-  2.47
point    17 -    18, DG -37.14 +/-  3.61
point    18 -    19, DG -18.76 +/-  6.09
point    19 -    20, DG 45.67 +/-  6.72

total    0 -    20, DG 348.47 +/- 23.20

正常吗，如何修改使得结果稳定

作者
Author: Huschein 时间: 2026-3-5 13:03
1. 增加窗口 2. 增加模拟时间 3. 调整lambda系数直接找一些基于gmx的文献看他的系数一般都不是这种固定间隔的 4. vdw和ele分开取消先取消ele 最后其实只要重叠矩阵质量可以标准差没有关系

欢迎光临计算化学公社 (http://bbs.keinsci.com/)