Hi Andrey,

The initial parameter file *.param.init for the B-spline potential function
contains break-points in the first column and knot values in the second
column.

As you observed in csg-tutorial/spce/re/ example, the initial values for
the knot values can be set to the PMF obtained by inverting the target
RDFs, i.e., c_i = -kTlog(g(r_i)). I have found this approach of
initializing knot values work well in my RE coarse-graining experiments.

Best,
Sikandar

On Wed, Jul 18, 2018 at 4:42 PM 'Andrey Brukhno' via votca <
[email protected]> wrote:

> Well, to reduce the burden on the developers' answering (probably obvious
> for them) questions, here is what I discovered by simply plotting the
> CG-CG.param.init file from the tutorial in csg-tutorials/spce/re/
>
> The CG-CG.param.init file appears to contain the -RT\ln(g_{target}(r))
> data sampled at the spline knots, i.e. the first column is for the knot
> values (r) and the second column is the corresponding PMF in kJ/mol - see
> the attached plot. I am not sure what is the origin for the slight
> discrepancies that are seen in the graph (resampling artefacts?), but it
> seems that the mystery about this file has been cleared up.
>
> Andrey
>
> On Wednesday, July 18, 2018 at 10:23:27 PM UTC+1, Andrey Brukhno wrote:
>>
>> Wow, nobody can answer this straightaway? - I wonder: does this mean that
>> no non-associate of the VOTCA team ever used RE method for their actual
>> problem (i.e. not a tutorial case)?
>>
>> Anyway, may I ask those on the team now: in what relation VOTCA's
>> B-spline definition is with the corresponding Sci-Py class
>> (scipy.interpolate.BSpline)?
>>
>>
>> https://docs.scipy.org/doc/scipy/reference/generated/scipy.interpolate.BSpline.html
>>
>> I also have another question: what is the actual format of the
>> .param.init file? - it is far from being clear in the manual, as the file
>> itself is never described.
>>
>> Thanks again.
>>
>> Andrey
>>
>> On Wednesday, July 18, 2018 at 6:51:44 PM UTC+1, Andrey Brukhno wrote:
>>>
>>> Hello,
>>>
>>> In an attempt to compare some data obtained with IBI to RE, I am trying
>>> to start a relative entropy iteration for the same system. However, I do
>>> not know how to generate the initial parameters (aka CG-CG.param.init file)
>>> for my system (which would comprise 10 different interaction types).
>>>
>>> Initially I thought VOTCA would be able to generate the necessary
>>> *.param.init files based on my initial potentials provided as *.pot.in
>>> (the IBI results) and/or distributions (*.dist.tgt). But it seems not to be
>>> the case, as it complaints in the inverse.log about missing *.param.init
>>> files.
>>>
>>> Can anyone suggest a workaround? - I googled and searched here but did
>>> not find anything regrading a practical way of creating those files (nor in
>>> the manual).
>>>
>>> Please note: my question is very practical - I know how fit numerical
>>> data with regular cubic splines, say in Grace, but VOTCA is using B-splines
>>> which are more general (which is good!) but also more involved, so I guess
>>> the coefficients of cubic splines cannot be used as initial "guess" for the
>>> RE iteration.
>>>
>>> I essentially need to know what software could do that - ?
>>>
>>> Thanks in advance!
>>>
>>> Andrey
>>>
>> --
> You received this message because you are subscribed to the Google Groups
> "votca" group.
> To unsubscribe from this group and stop receiving emails from it, send an
> email to [email protected].
> To post to this group, send email to [email protected].
> Visit this group at https://groups.google.com/group/votca.
> For more options, visit https://groups.google.com/d/optout.
>

-- 
You received this message because you are subscribed to the Google Groups 
"votca" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to [email protected].
To post to this group, send email to [email protected].
Visit this group at https://groups.google.com/group/votca.
For more options, visit https://groups.google.com/d/optout.

Reply via email to