Subject: Re: sparky-mars
From: ethernet3399
Date: Dec 23, 2005

Previous: 60

Dear Vitaliy Gorbatyuk,

I have tryed the rp command in Sparky to read in the sparky_CO-
1.out file. It works well, it can creat peak according the chemical
shifts.

I think we dont need to keep two sets of sparky project when run
Mars iteratively. Heres my suggestion:
Firstly, we can assign the peaks with pseduoresidue name and atom
name(CA,CA-1,CB, CB-1,CO,CO-1 et al.), then run Mars for the first
time.
When Mars finished, we can open the project(such as the spectrum
hncacb), delete all the peak, then rp to read in
the sparky_CA.out, sparky_CA-1.out, sparky_CB.out, sparky_CB-
1.out files in and create peaks.
I know in sparky_CA-1.out sparky_CB-1.out files the atom CA-1
is Ca, and CB-1 is Cb. This nomenclature cannot be recognized
by Mars for second run. So we can use the sparky rr command, and
rename the Ca to CA-1, Cb to CB-1, leave the residue(group)
name unchanged.
Of course we can extract the strip to check the Mars results in
sparky. We can modify the assignment if needed. Then we can output
the assignment table by tb command, and prepare to run Mars for
second time.

By the way, we can save the sparky assignment file *.save to
another name periodly. Thus if we do some thing wrong, we can just
open a previous *.save to correct the wrong assignment. *.proj
file is the same. These files are just ascii file with some texts.
Its small in size. To Keep them for several copies wouldnt take
much disk space. The spectrum files *.ucsf are just one copy from
begin to end.

May this help!

Tieying Zhang




--- In nmr_sparky@yahoogroups.com , ibmvg vitaliy_gor@h... wrote:

Thank you guys for your suggestions and time. I thought that MARS
developers meant a way which I could not figure out from their
paper.

Anyway, what about deleting a Sparky project? Yes, it takes a
little
of the disk space, but it may create enermous files and it makes it
difficult to find a certain file.

Cheers,
Vitaliy

--- In nmr_sparky@yahoogroups.com , ethernet3399 tyzhang@s...
wrote:

Dear Vitaliy Gorbatyuk,
I encountered the situation you mention.

I didnt read the sparky_CO-1.out into sparky HNCO spectrum.

While I notice the MARS program is written with awk language, so
you
can adjust the source code to make it output the CO-1, not
the Co. Can it fix your problem?

By the way, I agree with what Josh Ward says. I just rename the
residues name, such as PR1--A35, while leave the CO-1
atom
name unchanged. After all the MARS iteration finished, then we
can
rename the CO-1 atom name to T34 CO.

Of course, if youd like to change the MARS source code, please
let
the MARSs author know.

Or if you have any problem to rewrite the MARS awk source code,
we
can exchange idea about that further.

Tieying Zhang
at SIBS


--- In nmr_sparky@yahoogroups.com , Vitaliy Gorbatyuk
vitaliy_gor@h... wrote:

Thank you Josh.
I think what I was asking is how people do the input files for
the
second,
third, etc. run of Mars.
After the first run we get these files: sparky_CO.out,
sparky_CO-
1.out, etc.
As it is suggested in the MARS paper, I read in the peaks from
sparky_CO-1.out into HNCO spectrum with Sparky rp command.
Similarly, I
read other sparky_*.out files into the corresponding specta.
Then,
I use a
strip tool to build a sequence of strips and visually analyze
the
sequential
assignment obtained in the first run of MARS. After the
analysis I
want to
run MARS for the second time. But the name format of the peaks
is
CO,Co,CA,Ca, and so on. Thus, I have to go to the spectra
where I
have
manually done peak picking of pseudoresidues and where the
peak
names are
CO, CO-1, CA, CA-1, and so on, and perform corrections there:
add
news
peaks, delete wrong peaks, correct typos, which all this was
found
during
the analysis of strips. After corrections, I again save a peak
table with
tb command, modify fix_con.tab, fix_ass.tab and run MARS the
second time.

Thus, basically I have two sets of the same triple resonance
spectra in
sparky format (two sparky projects if you wish). One was used
to
do peak
picking and assigning the spin systems. Another was used to
read
in peaks
from MARS sparky_*.out files and to verify/correct the MARS
sequential
assignment and my mistakes/typos. I consider this approach a
bit
cumbersome
and wonder if someone has a better one.

In this regard, did anyone think on how to delete a Sparky
project
without
going in all Sparky directories, finding proper files and
delete
them
manually?

Cheers,
Vitaliy


From: Josh Ward wardjm@p...
Reply-To: nmr_sparky@yahoogroups.com
To: nmr_sparky@yahoogroups.com
Subject: Re: [nmr_sparky] sparky-mars
Date: Mon, 19 Dec 2005 10:06:11 -0500
I am not exactly sure what you find so weird in the MARS
formats,
but
here is the general strategy that I use when running
assignments
with MARS.

I first manually label pseudoresidues in sparky using 15N-
HSQC,
HNCACB,
and CBCA(CO)NH (I have not performed CO assignments yet for my
protein.) I label each pseudoresidue, giving a number for the
group, N
and H for the backbone amide resonances, and CA, CA-1, CB, and
CB-1
carbon assignments.

I then save the assignment table after this [Peak -
Assignment -
assignment table (at)] and feed this table into MARS as input.

After running MARS, I print out the assignment table and the
MARS
output
file assignment_AAs.out and use the hardcopies at the computer
to
go
through and manually verify the MARS assignments.

After verification I use the rename resonances (rr) facility
to
change
the group names to resonances (ie. 1 -- A35, 2 -- E26, etc),
and
the
CA-1 and CB-1 atoms to match the appropriate i-1 residue. I
admit,
this
process is tedious, and I am looking for a way to automate it
(I
would
like to write a python extension to automate the renaming
process
but
havent yet found the time.)

Anyway...my two cents. :) Hopefully it makes sense, or helps,
or at
least shows you you might not be doing it the hardest way. ;)

ibmvg wrote:

Hello,

I wonder if somebody could share their way of working with
Sparky-
MARS
tandem on the backbone assignment. I do not like how I do it.
Making
the input for MARS is straightforward, however tedious. But
MARS
output
with all these Co, Ca, etc. seems to be weird since it
differs
from
both MARS input and IUPAC format. Thus, in addition to a set
of
spectra
where we pick NMR peaks and give them names in the format of
MARS
input, one has to also have a set of all spectra for reading
in
the
MARS output. This set of spectra with peaks from MARS output
I
use for
manual verification of MARS assignment. Then, I use the
first
set of
spectra with CO, CO-1, etc. format to correct/add those
mistakes/peaks
which I found in the second set of spectra using MARS output
result. It
seems to me a cumbersome way. Maybe I am missing some neat
trick
which
could significantly simplify the job.

ANY ideas/suggestions are welcome :)!

Thanks & Merry Xmas!







Yahoo! Groups Links









--
Josh Ward
Graduate Research Assistant
Purdue University
Department of Medicinal Chemistry and Molecular Pharamacology
Lily Hall of Life Sciences
Phone: (765) 494-2191





Yahoo! Groups Links