From: Jérôme Hénin (jerome.henin_at_ibpc.fr)
Date: Tue Feb 07 2017 - 10:43:53 CST
Dear Lucas,
Did the simulation run at all? There is a large singularity around RMSD=0,
so you cannot start the simulation from one of the reference states, it
will blow up. Actually the safest is to exclude the region of very low RMSD
from the ABF calculation (maybe below 0.2 A or something).
Jerome
On 7 February 2017 at 16:28, Jessen Lucas <jessenlucas0_at_gmail.com> wrote:
> Dear Jerome,
>
>
> I am sorry to disturb you again, I tried to use ABF in NAMD to run
> deca-alanine in your paper using three RMSD collective variables, however
> it did not work, the .grad file show me all 0. Could you help me to point
> out the problem, or may I have your files about this case?
>
> The following is my namd conf file and colvar in file
>
> # NUMBER OF MD-STEPS
>
> numsteps 5000000
>
>
> # TOPOLOGY
>
> structure deca-ala.psf
>
>
> # FORCE FIELD
>
> parameters par_all22_prot.inp
> paraTypeCharmm on
>
>
> # 1-4 TERMs
>
> exclude scaled1-4
> 1-4scaling 1.0
>
>
> # INPUT FILES
>
> coordinates deca-ala.pdb
>
> temperature 300.0
>
>
> # OUTPUT FILES
>
> binaryoutput no
> binaryrestart yes
>
> outputname abf_00_0
> restartname abf_00
>
>
> # DCD FILE
>
> dcdFile abf_00.dcd
>
>
> # FREQUENCY FOR DUMPING OUTPUT DATA
>
> outputenergies 1000
> outputtiming 1000
> outputpressure 1000
> restartfreq 1000
> XSTFreq 1000
> dcdFreq 1000
>
>
> # CUT-OFFs
>
> hgroupcutoff 2.8
> switching on
> switchdist 9.0
> cutoff 11.0
> pairlistdist 12.0
>
>
> # CONSTANT-T
>
> langevin on
> langevintemp 300.0
> langevindamping 10.0
>
>
> # MULTIPLE TIME-STEP PROPAGATOR
>
> timestep 0.5
>
>
> # SHAKE/RATTLE
>
> rigidbonds none
>
>
> # PARALLELISM
>
> stepspercycle 20
> splitpatch hydrogen
> margin 1.0
>
>
> # ABF SECTION
>
> colvars on
> colvarsConfig Distance.in
> ##################################################
>
> colvarsTrajFrequency 500
> colvarsRestartFrequency 500
>
>
> colvar {
> name rmsdA
>
> outputTotalForce on
> # outputAppliedForce on
> width 1.0
>
> lowerBoundary 0.0
> upperBoundary 5.0
>
> lowerWallConstant 100.0
> upperWallConstant 100.0
>
> rmsd {
> atoms {
> atomNumbers 4 14 24 34 44 54 64 74 84 99
> }
> refpositionsfile helix.pdb
> }
> }
>
>
> colvar {
> name rmsdB
>
> outputTotalForce on
> # outputAppliedForce on
> width 1.0
>
> lowerBoundary 0.0
> upperBoundary 5.0
>
> lowerWallConstant 100.0
> upperWallConstant 100.0
>
> rmsd {
> atoms {
> atomNumbers 10 20 30 40 50 60 70 80 90 92
> }
> refpositionsfile 310turn.pdb
> }
> }
>
>
> colvar {
> name rmsdC
>
> outputTotalForce on
> # outputAppliedForce on
> width 1.0
>
> lowerBoundary 0.0
> upperBoundary 5.0
>
> lowerWallConstant 100.0
> upperWallConstant 100.0
>
> rmsd {
> atoms {
> atomNumbers 12 22 32 42 52 62 72 82 97
> }
> refpositionsfile omega.pdb
> }
> }
>
>
> abf {
> colvars rmsdA rmsdB rmsdC
> fullSamples 100
> }
>
>
> Is there something wrong in this files? I just follow the tutorial.
>
> Best regards,
> Lucas
>
>
> 2017-02-01 21:58 GMT+08:00 Jérôme Hénin <jerome.henin_at_ibpc.fr>:
>
>> Yes, it is written for arbitrary dimension.
>>
>> Jerome
>>
>> On 1 February 2017 at 14:40, Jessen Lucas <jessenlucas0_at_gmail.com> wrote:
>>
>>> Thank you, so the abf_integrate works for high dimension?
>>>
>>> Lucas
>>>
>>> 2017-02-01 21:20 GMT+08:00 Jérôme Hénin <jerome.henin_at_ibpc.fr>:
>>>
>>>> Yes, except that you dont want those lines!
>>>> centerReference no
>>>> rotateReference no
>>>>
>>>> Once you have integrated the free energy gradient with abf_integrate,
>>>> you get the PMF in a plain text file. You can plot it, but you can also
>>>> open it in a text editor, and you'll find individual values in there.
>>>>
>>>> Jerome
>>>>
>>>>
>>>> On 1 February 2017 at 14:16, Jessen Lucas <jessenlucas0_at_gmail.com>
>>>> wrote:
>>>>
>>>>> Thanks Jerome, so it should be like the following?
>>>>>
>>>>> colvar {
>>>>> name rmsdA
>>>>>
>>>>> width 1.0
>>>>>
>>>>> lowerBoundary 0.0
>>>>> upperBoundary 5.0
>>>>>
>>>>> lowerWallConstant 100.0
>>>>> upperWallConstant 100.0
>>>>>
>>>>> rmsd {
>>>>> atoms {
>>>>> atomNumbersRange 1-104
>>>>> centerReference no
>>>>> rotateReference no
>>>>> }
>>>>> refpositionsfile ../Common/omega-ideal.pdb
>>>>> }
>>>>> }
>>>>>
>>>>> And I saw the ABF output the whole free energy landscape, is there
>>>>> anyway to get one point free energy? for example, the free energy value at
>>>>> rmsdA=3, rmsdB=4?
>>>>>
>>>>> Thank you very much for your kindly reply.
>>>>> Lucas
>>>>>
>>>>> 2017-02-01 21:07 GMT+08:00 Jérôme Hénin <jerome.henin_at_ibpc.fr>:
>>>>>
>>>>>> Your syntax for the RMSD components is not correct. The "atoms" block
>>>>>> needs to contain the definition of which atoms are involved. The
>>>>>> "refpositionsfile" is a parameter of the rmsd function, so it should be
>>>>>> outside that block.
>>>>>>
>>>>>> Best,
>>>>>> Jerome
>>>>>>
>>>>>> On 31 January 2017 at 17:04, Jessen Lucas <jessenlucas0_at_gmail.com>
>>>>>> wrote:
>>>>>>
>>>>>>> Dear Jerome,
>>>>>>>
>>>>>>> I found that in github colvars
>>>>>>> <https://urldefense.proofpoint.com/v2/url?u=https-3A__github.com_colvars_colvars&d=DwMFaQ&c=8hUWFZcy2Z-Za5rBPlktOQ&r=6cQ04kjNHUuxT8SDe1wb0TtPDsnWRp4T4iwoYRLFpsY&m=i83dq7tEuia45zFsvze217Dl3QR4ygres_qi16tY7xU&s=BPqNf54jff0GZnsV9qDB5_TyL-vmmzxJCxGv7DXMzzA&e=>
>>>>>>> /namd
>>>>>>> <https://urldefense.proofpoint.com/v2/url?u=https-3A__github.com_colvars_colvars_tree_master_namd&d=DwMFaQ&c=8hUWFZcy2Z-Za5rBPlktOQ&r=6cQ04kjNHUuxT8SDe1wb0TtPDsnWRp4T4iwoYRLFpsY&m=i83dq7tEuia45zFsvze217Dl3QR4ygres_qi16tY7xU&s=z9SA_KG80pGsEuFAgT1_Wh_l3U_Bln-cHm8FGtvNHU4&e=>
>>>>>>> /tests
>>>>>>> <https://urldefense.proofpoint.com/v2/url?u=https-3A__github.com_colvars_colvars_tree_master_namd_tests&d=DwMFaQ&c=8hUWFZcy2Z-Za5rBPlktOQ&r=6cQ04kjNHUuxT8SDe1wb0TtPDsnWRp4T4iwoYRLFpsY&m=i83dq7tEuia45zFsvze217Dl3QR4ygres_qi16tY7xU&s=VRHQSpFoobNIjYT4dF6jkIJxsyLxy3cJqTTlKI_eazg&e=>
>>>>>>> /library
>>>>>>> <https://urldefense.proofpoint.com/v2/url?u=https-3A__github.com_colvars_colvars_tree_master_namd_tests_library&d=DwMFaQ&c=8hUWFZcy2Z-Za5rBPlktOQ&r=6cQ04kjNHUuxT8SDe1wb0TtPDsnWRp4T4iwoYRLFpsY&m=i83dq7tEuia45zFsvze217Dl3QR4ygres_qi16tY7xU&s=dN7xFNniZSYnuJj9_W8kpjcJsnxM-ulZ5pOI6r-yTk4&e=>
>>>>>>> /001_10ala_RMSD
>>>>>>> <https://urldefense.proofpoint.com/v2/url?u=https-3A__github.com_colvars_colvars_tree_master_namd_tests_library_001-5F10ala-5FRMSD&d=DwMFaQ&c=8hUWFZcy2Z-Za5rBPlktOQ&r=6cQ04kjNHUuxT8SDe1wb0TtPDsnWRp4T4iwoYRLFpsY&m=i83dq7tEuia45zFsvze217Dl3QR4ygres_qi16tY7xU&s=_vxHuW7Gq0XDMSTpKyG2GUzlQCaNaHVHl9lyaKBbyko&e=>
>>>>>>> /, you use 5 rmsd as collective variables. However when I tried to
>>>>>>> change those collective variables to the three rmsd collective variables as
>>>>>>> in your paper "Exploring Multidimensional Free Energy Landscapes
>>>>>>> Using Time-Dependent Biases on Collective Variables"?
>>>>>>>
>>>>>>> Is the following colvar config correct?
>>>>>>>
>>>>>>> colvar {
>>>>>>> name rmsdA
>>>>>>>
>>>>>>> outputTotalForce on
>>>>>>>
>>>>>>> width 1.0
>>>>>>>
>>>>>>> lowerBoundary 0.0
>>>>>>> upperBoundary 5.0
>>>>>>>
>>>>>>> lowerWallConstant 100.0
>>>>>>> upperWallConstant 100.0
>>>>>>>
>>>>>>> rmsd {
>>>>>>> atoms {
>>>>>>> refpositionsfile ../Common/310turn-ideal.pdb
>>>>>>> }
>>>>>>> }
>>>>>>> }
>>>>>>>
>>>>>>> colvar {
>>>>>>> name rmsdB
>>>>>>>
>>>>>>> outputTotalForce on
>>>>>>>
>>>>>>> width 1.0
>>>>>>>
>>>>>>> lowerBoundary 0.0
>>>>>>> upperBoundary 5.0
>>>>>>>
>>>>>>> lowerWallConstant 100.0
>>>>>>> upperWallConstant 100.0
>>>>>>>
>>>>>>> rmsd {
>>>>>>> atoms {
>>>>>>> refpositionsfile ../Common/helix-ideal.pdb
>>>>>>> }
>>>>>>> }
>>>>>>> }
>>>>>>>
>>>>>>> colvar {
>>>>>>> name rmsdC
>>>>>>>
>>>>>>> outputTotalForce on
>>>>>>>
>>>>>>> width 1.0
>>>>>>>
>>>>>>> lowerBoundary 0.0
>>>>>>> upperBoundary 5.0
>>>>>>>
>>>>>>> lowerWallConstant 100.0
>>>>>>> upperWallConstant 100.0
>>>>>>>
>>>>>>> rmsd {
>>>>>>> atoms {
>>>>>>> refpositionsfile ../Common/omega-ideal.pdb
>>>>>>> }
>>>>>>> }
>>>>>>> }
>>>>>>>
>>>>>>> I do not why it did not work when I do a small change? After that
>>>>>>> instead of using colvar module of NAMD, I use tcl script I could calculate
>>>>>>> the rmsd, but I do not know how to calculate the free energy at the point.
>>>>>>>
>>>>>>> Could you give me a suggestion?
>>>>>>> Thanks in advance.
>>>>>>>
>>>>>>> Lucas
>>>>>>>
>>>>>>>
>>>>>>> 2017-01-14 7:48 GMT+08:00 Jessen Lucas <jessenlucas0_at_gmail.com>:
>>>>>>>
>>>>>>>> Thank you, dear Jerome. :)
>>>>>>>>
>>>>>>>> 2017-01-14 0:31 GMT+08:00 Jérôme Hénin <jerome.henin_at_ibpc.fr>:
>>>>>>>>
>>>>>>>>> Dear Lucas,
>>>>>>>>>
>>>>>>>>> Here are the files in question.
>>>>>>>>>
>>>>>>>>> Jerome
>>>>>>>>>
>>>>>>>>> On 13 January 2017 at 15:55, Jessen Lucas <jessenlucas0_at_gmail.com>
>>>>>>>>> wrote:
>>>>>>>>>
>>>>>>>>>> Dear Jerome,
>>>>>>>>>>
>>>>>>>>>> Yes, I want to use the same system, may I have your files these
>>>>>>>>>> three conformations, since I want to get the same result as yours,
>>>>>>>>>> otherwise it is hard to convince myself. Thanks a lot.
>>>>>>>>>>
>>>>>>>>>> Best regards,
>>>>>>>>>> Lucas
>>>>>>>>>>
>>>>>>>>>> 2017-01-13 16:57 GMT+08:00 Jérôme Hénin <jerome.henin_at_ibpc.fr>:
>>>>>>>>>>
>>>>>>>>>>> Hi Lucas,
>>>>>>>>>>>
>>>>>>>>>>> For that work, the reference conformations were not strictly
>>>>>>>>>>> speaking "ideal" (if I wrote that, it was a poor choice of words). They
>>>>>>>>>>> were local energy minima of that particular peptide under those particular
>>>>>>>>>>> conditions, and I selected them from ABF trajectories using different
>>>>>>>>>>> coordinates.
>>>>>>>>>>>
>>>>>>>>>>> If you are simulating the very same system, I will gladly share
>>>>>>>>>>> the files I used (perhaps off-list) - however they don't have any general
>>>>>>>>>>> value: some of them are rather idiosyncratic of that peptide.
>>>>>>>>>>>
>>>>>>>>>>> Best,
>>>>>>>>>>> Jerome
>>>>>>>>>>>
>>>>>>>>>>> On 13 January 2017 at 03:09, Jessen Lucas <
>>>>>>>>>>> jessenlucas0_at_gmail.com> wrote:
>>>>>>>>>>>
>>>>>>>>>>>> Dear all,
>>>>>>>>>>>>
>>>>>>>>>>>> I am reading "Exploring Multidimensional Free Energy Landscapes
>>>>>>>>>>>> Using Time-Dependent Biases on Collective Variables", in there, I want to
>>>>>>>>>>>> plot the free energy landscape, I saw there use three collective variables,
>>>>>>>>>>>> the rmsd with respective to the ideal alpha-helical, ω-shaped
>>>>>>>>>>>> γ-turn, and turn/310-helix conformations, respectively. My question is how
>>>>>>>>>>>> to set this three collective variables in NAMD? I do not know what the
>>>>>>>>>>>> conformations of ideal alpha-helical, ω-shaped γ-turn, and
>>>>>>>>>>>> turn/310-helix.
>>>>>>>>>>>>
>>>>>>>>>>>> Thanks for any kindly suggestion.
>>>>>>>>>>>> Lucas
>>>>>>>>>>>>
>>>>>>>>>>>
>>>>>>>>>>>
>>>>>>>>>>
>>>>>>>>>
>>>>>>>>
>>>>>>>
>>>>>>
>>>>>
>>>>
>>>
>>
>
This archive was generated by hypermail 2.1.6 : Mon Dec 31 2018 - 23:20:05 CST