To Do List


Currently the To Do lists of different projects can be found from the blog post beginning with the project identification (e.g. NMRlipidsIVb) and from manuscript drafts and issue trackers in the repositories in the NMRlipids GitHub.

Comments here refer to obsolete ToDo lists which were published in this page. The comments are saved due to their scientific content.

7 comments:

  1. I have now started simulations with different NaCl concentrations using the CHARMM. I will also start simulations with different CaCl_2 concentrations.

    ReplyDelete
  2. Hello,

    Here are the results for the hybrid FF:CHARMM36+Berger (see http://pubs.acs.org/doi/abs/10.1021/jp800687p). The system consists of 1 bilayer (128 lipids) with 30 waters/lipid.

    DLPC@323K
    beta -0.05974 -0.08365
    alpha -8.24E-4 0.03133
    g3R -0.23074
    g3S -0.21543
    g2 -0.18188
    g1R -0.15811
    g1S -0.02211

    DPPC@298K
    beta -0.07576 -0.08963
    alpha 0.01124 0.05785
    g3R -0.24477
    g3S -0.28501
    g2 -0.31129
    g1R 0.07088
    g1S -0.30659

    ReplyDelete
  3. Thanks! I have a couple of questions:

    1) You mention CHARMM36+Berger combination, however CHARMM27 is used in the paper you have linked. Which one did you use?

    2) Which lipid do you refer with DLPC? I have seen that it is used for the dilinoleoyl-PC and for the dilauroyl-PC.

    3) You seem to have the DPPC in low temperature. Presumably it is in the gel phase?

    ReplyDelete
    Replies
    1. 1) We took FF from http://pubs.acs.org/doi/abs/10.1021/jp800687p and replaced CHARMM27 by CHARMM36. This corresponds to the paper you cited ( 10.1021/jp410344g)
      2) CH3-(CH2)10-COOH - lauric
      3) you're right, gel phase.

      I must admit that the product run was only 1 ns. I am now recomputing all values for long simulations.

      Delete
    2. I have updated the results (average over 20 ns)

      DLPC@323K (liquid phase)
      beta -0.06525 -0.06584
      alpha 0.03839 0.02978
      g3R -0.24805
      g3S -0.21785
      g2 -0.16463
      g1R -0.15191
      g1s -0.04938


      DPPC@298K (gel phase):
      beta -0.09415 -0.06091
      alpha 0.02928 0.04799
      g3R -0.24558
      g3S -0.28782
      g2 -0.31054
      g1R 0.04387
      g1s -0.31255

      Delete

  4. Hi !
    I have been working with Alexandru Botan during his post-doc in Lyon, and I want to precise the model in the contribution by Alexandru Botan, after a discussion with Luca Monticelli and Markus Miettinen last week in Lyon.

    The model UA-AA we have used here is actually different both from the
    UA-charmm27 by Henin et al (2008) and from the UA-charmm36 by Lee et al (2014). We have taken the topology by Henin et al., and used the charmm 36 parameters for the heads. Therefore, we should not pretend our model is exactly one of these.

    To answer one question by Luca, it is not a huge problem to keep
    the AA-AU dihedrals from Henin et al, based on charmm27, despite the
    change of charges on the head for charmm36. We have computed the dihedral distributions of beta3, beta4, gamma3, gamma4. They are qualitatively similar in our AA-AU model and in the AA-charmm36. A difference exists in the amount of trans and gauche in beta4 and gamma4. The beta3 and gamma3 distribution match very satisfactorily in the AA and our AA-UA model.

    As these details are not the point of your article, it is
    probably more clear to replace the data from our home-made model
    by data from simulations using the CHARMM-36-UA by Lee et al (2014).

    Here they are : I have used namd2.10 and calculated
    the order parameters for such a the CHARMM-36-UA by Lee et al (2014).
    72 DPPC + 2189 water, 323K, NPg=0T (liquid phase)
    30 ns equilibration + 50 ns analysis, total time 80ns.
    (initial state taken from J. Klauda web page, charmm36-AA configuration).
    I have used the topology and parameter files from the SI of the article.
    Minor modifications to transfer the charmm format to NAMD format.

    Here are the mean values
    (no analysis of correlation time,
    I have analyzed about 2600 configurations, separated by 20 ps):

    beta1 -0.067
    beta2 -0.062
    alpha1 0.036
    alpha2 0.033
    g31 -0.239
    g32 -0.258
    g2 -0.195
    g11 -0.172
    g12 -0.047



    So, for the article to be submitted very soon,
    I see two solutions : either you keep alexandru's results and explain a
    bit more the model in the SI; or you add/replace by these results from charmm36-AU- Lee's 2014 model.

    I can provide the input files and the final trajectory (NAMD2.10).
    If you want some dihedral distributions, for example for Fig. just tell me.

    Best regards,
    Claire

    ReplyDelete
    Replies
    1. Thank you for the contribution. I think that we could do the following: We keep the data delivered by Botan in Table 1 and in Fig. 2. However, we use the data from actual Lee et al. model in the Fig. 4 (ranking figure).

      We just need to decide how we call your "home-made" model parameters, and we also have to modify the simulation details accordingly.

      I have now added the results you delivered into github, but not in figure yet: https://github.com/NMRLipids/nmrlipids.blogspot.fi/blob/master/DATAreportediINblog/DPPC/CHARMM36UA-323K_blogged-25-3-15.dat

      Couple of questions:
      1) Is there difference in your approach and approach by Lee et al. to make CHARMM36-UA model, or is it just two different implementations of the same thing?

      2) You mentioned that you had some differences in beta4 and gamma4 gaughe/trans ratio between CHARMM-UA and CHARMM-AA. Did you see differences in order parameters near these dihedrals?

      Delete

Please sign in before writing your comment.