Re: [ESPResSo-users] protein self-aggregation, directional lennard jonnes

[Tristan Bereau]

Dear Sarah,

You need to set the PEPTIDEB_DIR environment variable to point to the
peptideB directory in your system. If, say, you use bash and have
saved the package in ~/peptideB, call the command:

export PEPTIDEB_DIR=~/peptideB

you can then check that the environment variable was stored correctly
by invoking:

echo $PEPTIDEB_DIR

making sure that it returns the directory of the package. Rather than
typing the first command every time you open a terminal, you can
automate it by saving it in your ~/.bashrc file.

(If you're using tcsh, have a look on the Internet on how to set
environment variable for that environment.)

Best,
Tristan

On Mon, Feb 4, 2013 at 9:11 AM, sarah mohamadinegad <address@hidden> wrote:
> Dear Tristan
> I hope I can cc it this time, the last time I tried to cc my e-mail but I
> couldn't. Sorry!
>
> Very good news,
> So the point is that *relative* PID is important not the PID itself! Thanks
> Tristan.
>
> I have one more question.
> As you referred me to peptideB, I take a look and find it exactly what I
> need. But the installation guide is not clear enough to me. When I want to
> run one of your sample code "peptide.tcl", using this command:
>
> Espresso peptidebuilder.tcl peptide.tcl -espresso
>
> I faced to an error:
>
> Error. Can't find the PEPTIDEB_DIR environment variable.
>
> I know the mistake is with
>
> export PEPTIDEB_DIR=path/peptideB
>
> but I don't know exactly how and maybe where to fix it! Would you please
> help me with this or refer me to a more detailed installation guide?
>
> Thanks in advance,
> Sarah
>
>
>
>
>
>
> ________________________________
> From: Tristan Bereau <address@hidden>
> To: sarah mohamadinegad <address@hidden>
> Cc: "address@hidden" <address@hidden>
> Sent: Monday, February 4, 2013 11:19 AM
>
> Subject: Re: [ESPResSo-users] protein self-aggregation, directional lennard
> jonnes
>
> Dear Sarah,
>
> (I'm cc-ing this to the mailing list.)
>
> The interaction only needs to be set once for a pair of atom types, so
> there would only be one call to that command. The last 4 numbers in
> the 'inter' command refer to the *relative* position of neighboring
> atoms along the backbone chain. They allow to define the angles for
> the directionality of the interaction. Here's an example from the code
> I referred to in my last email:
>
> inter 1 2 lj-angle $eps $sigma $cut 1 -1 1 -2
>
> that describes an interaction between types 1 (i.e., N) and 2 (i.e.,
> C), where the direction of the amide group is given by the three atoms
> [N-1] N and [N+1] (consecutive along the chain), and the direction of
> the carbonyl group by [C-2] C and [C+1] (thus the "-2" at the end of
> the command. All in all, the four numbers you need to type in at the
> end of the command depend on the arrangement of beads (i.e., PIDs in
> ESPResSo) you use to create your backbone chain.
>
> Best,
> Tristan
>
> On Mon, Feb 4, 2013 at 7:22 AM, sarah mohamadinegad <address@hidden>
> wrote:
>> Dear Tristan,
>>
>> Thanks for your kind answer.
>>
>> Maybe describing my trouble using an example will help:
>>
>> Suppose a 200aa protein and I assign type=1 to all my C atoms and type=2
>> to
>> all N atoms. Now I need to have a hydrogen bond between a C atom with pid
>> 15
>> and N atom with pid 148, (type=1,pid=15 -> type=2,pid=148), So using
>> directional LJ, I write
>>
>> inter 1 2 lj-angle $eps $sigma $cut 14 16 147 149
>>
>> Just here I have a question! this line set an interaction between all C
>> atoms(type=1) and all N atoms(type=2) present in my box or only my desired
>> atoms?
>>
>> If I want to set hydrogen bond between all C atoms and all N atoms, what
>> should I write in place of the four last pids (14 16 147 149)?
>>
>> Thanks in advance,
>> Sarah
>>
>>
>>
>>
>>
>> ________________________________
>> From: Tristan Bereau <address@hidden>
>> To: sarah mohamadinegad <address@hidden>
>> Cc: "address@hidden" <address@hidden>
>> Sent: Sunday, February 3, 2013 6:04 PM
>> Subject: Re: [ESPResSo-users] protein self-aggregation, directional
>> lennard
>> jonnes
>>
>> Dear Sarah,
>>
>> How you want to model a system ultimately depends on what you want to
>> recreate. However, atom types for protein backbones are generally
>> residue-independent, i.e., one set of parameters irrespective of the
>> residue. The interaction potential you refer to was used in this paper
>> to model CG peptides:
>>
>> http://jcp.aip.org/resource/1/jcpsa6/v130/i23/p235106_s1
>>
>> This model did use one atom type for the C atoms and another one for
>> the N atoms. See the following link for an ESPResSo implementation of
>> the CG model:
>>
>> https://github.com/tbereau/peptideB
>>
>> Best,
>> Tristan
>>
>> On Sun, Feb 3, 2013 at 2:59 PM, sarah mohamadinegad <address@hidden>
>> wrote:
>>> Dear users and developers,
>>>
>>> I am going to model a protein (almost large >200 aa) self-aggregation by
>>> using ESPResSo. To set hydrogen bond between amino acids of protein, I am
>>> going to use directional lennard jonnes (lj-angle).
>>>
>>> As I should consider hydrogen bond between N-H group of every amino acid
>>> with C-O group of other amino acids, do I have to assign different types
>>> to
>>> N and C groups of different amino acids?
>>> For example:
>>> aa1  : N(type = 1) C(type = 2)
>>> aa2  : N(type = 3) C(type = 4)
>>> aa3  : N(type = 5) C(type = 6)
>>> aa4  : N(type = 7) C(type = 8)
>>> aa5  : N(type = 9) C(type = 10)
>>> .
>>> .
>>> .
>>> aa200: N(type = 399) C(type = 400)
>>>
>>> I would be happy if I can assign only one type (for example 1) to all of
>>> my
>>> N atoms and another one (for example 2) to all my C atoms, Is it possible
>>> for the case of my problem?
>>>
>>> Thanks in advance for your help,
>>> Sarah
>>>
>>>
>>
>>
>
>