## Re: linear prediction in N15 dimension

Expand Messages
• oh my! there are numerous mistakes in the scheme. why are you scaling the first points of the FID with -c .33 or -c .66? this will introduce a baseline
Message 1 of 2 , Jul 10, 2007
• 0 Attachment
oh my!

there are numerous mistakes in the scheme.

why are you scaling the first points of the FID with
-c .33 or -c .66? this will introduce a baseline distortion.

scale the first point with -c 0.5 when there is no first order
phase (i.e. no delay at the start of that dimension). leave the
first point unchanged otherwise (-c 1.0)

and, you are applying a baseline correction:

POLY -ord 1

but, you havent specified any baseline points. i,m not sure what
will happen in this case, probably nothing. but, you shouldnt need
any baseline correction here, especially if the first point scaling
has been set correctly.

now, most commonly, in a scheme where inverse processing is used,
the inverse scheme may result in distortions of either the
first or last point in the inverse-transform FID, or maybe even
both. this will distort the LP results. this is especially true for
mirror image LP. its safer to work with "LP -fb" at least until
you figure out the other aspects of your scheme ... see the

also, you have the wrong use of the "-inPlace" argument; this is
only used when the source and destination names are the same.
if you are "debugging" a processing scheme, its best not to
use in-place mode at all, for one reason being that you
can inspect the intermediate results.

also also, you have a zero-fill value hard-wired into your scheme
(ZF -size 128) ... this is a problem if your requested
size is less than 2x the original time-domain size.

if you want to fully explore the situation, first try LP
in one dimension only:

1. process the data as: X, then Y, then Z with "LP -fb"
check that the result is ok.

2. process the data as X, then Z, then Y with "LP -fb"
check that the result is ok

then, use your two-dimensional LP scheme, but save the
interferogram as an intermediate step, for example:

> xyz2pipe -in lp/test_hnco%03d.ft3 -z -verb \
> | nmrPipe -fn HT -auto \
> | nmrPipe -fn PS -inv -hdr \
> | nmrPipe -fn FT -inv \
> | nmrPipe -fn ZF -inv \
> | nmrPipe -fn SP -inv -hdr \
> | pipe2xyz -out lp/int%03d.dat

then, have a look at the interferogram, to see if either the
first points or last points are distorted. if the first
point is distorted, this is probably a first-point scaling
issue. if the last point is distorted, you can add an
"EXT" to extract all but the last point (i.e. discard the
bad last point) and then do LP.

hope this helps ...

--- In nmrpipe@yahoogroups.com, "jhhong7546" <croshong@...> wrote:
>
>
> Hi all
>
> This time I wanna have your opinion about the linear prediction
>
> Recently I got the HNCO and HNCACB spectrum. Both of them looks
>
> good when they were processed using just FT. but When I use
>
> Linear prediction in C13 and N15 dimension especially
>
> in N15 dimension spectrum became somewhat strange. for example
>
> some peak shape is severely distorted and can not be phased
properly
>
> I adjusted points which is predicted and some other parameter
>
> but it could not become good. Is there any kind of something like
>
> rule for applying linear prediction about for example in certain
>
> case adding how much ponit is suitbale and above that spectrum
>
> will became very bad? This is my processing script
>
> for HNCO
>
> xyz2pipe -in fid/test_hnco%03d.fid -x -verb \
> | nmrPipe -fn POLY -time \
> | nmrPipe -fn SP -off 0.5 -end 0.98 -pow 2 -c 1.0 \
> | nmrPipe -fn ZF -auto \
> | nmrPipe -fn FT \
> | nmrPipe -fn PS -p0 159.0 -p1 0.0 \
> | nmrPipe -fn EXT -x1 6.5PPM -xn 10.5PPM -di -sw \
> | pipe2xyz -out lp/test_hnco%03d.ft3 -x -ov
>
> xyz2pipe -in lp/test_hnco%03d.ft3 -z -verb \
> | nmrPipe -fn SP -off 0.5 -end 0.98 -pow 1.0 -c 0.67 \
> | nmrPipe -fn ZF -size 128 \
> | nmrPipe -fn FT -di \
> | pipe2xyz -out lp/fadddm_hnco%03d.ft3 -z -inPlace
>
> xyz2pipe -in lp/test_hnco%03d.ft3 -y -verb \
> | nmrPipe -fn LP -fb -pred 64 \
> | nmrPipe -fn SP -off 0.5 -end 0.98 -pow 2 -c 0.33 \
> | nmrPipe -fn ZF -auto \
> | nmrPipe -fn FT \
> #| nmrPipe -fn REV \
> | nmrPipe -fn PS -p0 7.0 -p1 0.0 -di \
> | nmrPipe -fn POLY -ord 1 \
> | pipe2xyz -out lp/test_hnco%03d.ft3 -y -inPlace
>
> xyz2pipe -in lp/test_hnco%03d.ft3 -z -verb \
> | nmrPipe -fn HT -auto \
> | nmrPipe -fn PS -inv -hdr \
> | nmrPipe -fn FT -inv \
> | nmrPipe -fn ZF -inv \
> | nmrPipe -fn SP -inv -hdr \
> | nmrPipe -fn LP -ps0-0 -pred 64 \
> | nmrPipe -fn SP -off 0.5 -end 0.98 -pow 1.0 -c 0.67 \
> | nmrPipe -fn ZF -size 128 \
> | nmrPipe -fn FT \
> | nmrPipe -fn PS -hdr -di \
> | pipe2xyz -out lp/test_hnco%03d.DAT -z -inPlace
>
> HNCACB processing were done nearly same scheme