[gmx-users] Clarity on TI free energy terms

[Nash, Anthony]

Hi Hannes,

Thanks for the reply. At the moment for a change in single amino acid in a
triplet (a pair of triplets, showing forward and reverse change) I am
settling with:

vdw_lambdas              = 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
mass_lambdas = 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
0.0 0.0 0.0 0.0 0.0 0.0
fep_lambdas =  0.0 0.1 0.1 0.2 0.2 0.3 0.3 0.4 0.4 0.5 0.5 0.6 0.6 0.7 0.7
0.8 0.8 0.9 0.9 1.0 1.0


free_energy              = yes
init_lambda_state        = 0
delta_lambda             = 0
calc_lambda_neighbors    = 1        ; only immediate neighboring windows


; Options for the decoupling
sc-alpha                 = 0.5
sc-coul                  = yes
sc-power                 = 1.0
sc-sigma                 = 0.3
couple-moltype           = protein
couple-lambda0           = vdw-q
couple-lambda1           = vdw-q
couple-intramol          = no
nstdhdl                  = 10



The  mass will be conserved (it is a full cycle in one system). Everything
else scales with feb_lambdas apart from vdw which will (I am guessing)
require more sampling. This is my first attempt, I don't expect to get a
true understanding of it yet :-)

Thanks again
Anthony





On 01/06/2016 12:43, "gromacs.org_gmx-users-bounces at maillist.sys.kth.se on
behalf of Hannes Loeffler"
<gromacs.org_gmx-users-bounces at maillist.sys.kth.se on behalf of
Hannes.Loeffler at stfc.ac.uk> wrote:

>
>Set the vector to all-zeroes (or ones).
>
>
>On Wed, 1 Jun 2016 09:47:59 +0000
>"Nash, Anthony" <a.nash at ucl.ac.uk> wrote:
>
>> Hi Hannes,
>> 
>> 
>> Many thanks for the reply. With regards to your final comment I
>> understand conserving mass in theory, but I am a little confused
>> regarding, "keep the mass-lambdas at one end-point as they can
>> interact badly with constraints". I am testing pmx on a two-molecule
>> one-system I.e., G-D2K-G and G-K2D-G in the same system. How ought I
>> define the mass-lambdas for this system? (nothing accurate, just an
>> example would be great)?
>> 
>> Thanks
>> Anthony
>> 
>> 
>> On 01/06/2016 09:55,
>> "gromacs.org_gmx-users-bounces at maillist.sys.kth.se on behalf of
>> Hannes Loeffler" <gromacs.org_gmx-users-bounces at maillist.sys.kth.se
>> on behalf of Hannes.Loeffler at stfc.ac.uk> wrote:
>> 
>> >On Wed, 1 Jun 2016 07:54:56 +0000
>> >"Nash, Anthony" <a.nash at ucl.ac.uk> wrote:
>> >
>> >> In the tutorial, charges are off in the topology and the
>> >> electrostatic coupling to lambda remains 0 throughout the 20
>> >> windows. I assume setting col_lambdas=0 0 0 Š was for that very
>> >> reason I.e., the charges were off? Could the charges not have been
>> >> left on and col_lambdas defined similar to vdw_lambdas?
>> >> (I understand that if charges remain constant, as vdw turns off,
>> >> the system will probably blow up as attraction brings molecules
>> >> infinitely close).
>> >
>> >Technically, Gromacs allows you to vary both vdW and Coulomb lambdas
>> >simultaneously because Gromacs can apply softcore potentials to both.
>> >In practice though it seems that many workers still prefer to
>> >separate the two terms from each other.
>> >
>> > 
>> >> If my transition is from a small molecule into a small molecule
>> >> e.g., G-D-G to G-K-D, (the PMX paper) should I define all three
>> >> lambdas: vdw_lambdas, col_lambdas and bonds_lambdas? Between
>> >> states A and B, VdW, charges and bonds are all changing.
>> >
>> >Lambda paths are only about separating the various force field terms
>> >from each other.  If you do not explicitly state any of those lambda
>> >vectors they will adopt they same lambdas as specified in
>> >fep-lambdas, see manual.  I do not see a reason why you would want
>> >to separate out the bonded terms as well.  They are subject to a
>> >linear transformation only anyway.
>> >
>> >What you may want to do is to keep the mass-lambdas at one end-point
>> >as they can interact badly with constraints.  In a proper
>> >thermodynamic cycle mass contributions must perfectly cancel.
>> >
>> >
>> >Cheers,
>> >Hannes.
>> >-- 
>> >Gromacs Users mailing list
>> >
>> >* Please search the archive at
>> >http://www.gromacs.org/Support/Mailing_Lists/GMX-Users_List before
>> >posting!
>> >
>> >* Can't post? Read http://www.gromacs.org/Support/Mailing_Lists
>> >
>> >* For (un)subscribe requests visit
>> >https://maillist.sys.kth.se/mailman/listinfo/gromacs.org_gmx-users or
>> >send a mail to gmx-users-request at gromacs.org.
>> 
>
>-- 
>Gromacs Users mailing list
>
>* Please search the archive at
>http://www.gromacs.org/Support/Mailing_Lists/GMX-Users_List before
>posting!
>
>* Can't post? Read http://www.gromacs.org/Support/Mailing_Lists
>
>* For (un)subscribe requests visit
>https://maillist.sys.kth.se/mailman/listinfo/gromacs.org_gmx-users or
>send a mail to gmx-users-request at gromacs.org.