Description Usage Arguments Details Value Author(s) See Also Examples
Core function to perform baseline correction on one metabolomics fingerprint.
1 2 |
x |
A numeric vector to be processed. |
wsize |
Window size. |
qtl |
A numeric value of for lower quantile probability. |
maxy |
A numeric value specifying y axis maximal value to be plotted. |
plotting |
A logical value indicating whether or not plotting. |
title |
A logical value indicating whether or not to show plot title. |
sampid |
Sample ID to be written in the title if both plotting and title are TRUE
(useful when |
The purpose of the baseline correction is to remove undesirable effects due
excess chemical noise. Possible consequences of baseline drift include the
possibility that the baseline may be discriminatory and that important
information may be obscured in areas affected by baseline problems. A simple
consensual approach consists in fitting a monotone local minimum curve to
each fingerprint. Basically, the fingerprint is divided into equally
spaced m/z intervals and a local minimum intensity value is returned as the
baseline estimate for this region. Finally, the whole fingerprint baseline is
computed by linear interpolation based on pairs made of the centre of the
interval and its corresponding local minima. Intervals (argument wsize
)
are in the order of 30-70 amu as a trade off between the removal of relevant
chemical (small interval) or estimation bias due to use of a larger interval.
Rather than using the minimum value of an interval, it is also judicious to
use the value corresponding to a low quantile (argument qtl
) to avoid
any spurious estimates due to zeros or abnormally low signals.
A list containing the following components:
x |
A numeric vector of the resulting fingerprint after baseline correction. |
bsl |
A numeric vector of the baseline intensities. |
David Enot dle@aber.ac.uk and Wanchang Lin wll@aber.ac.uk.
1 2 3 4 5 6 7 |
Add the following code to your website.
For more information on customizing the embed code, read Embedding Snippets.