This applet is a spin-off from my graduate thesis in biotechnology at the Universiteit Gent (RUG) in Ghent, Belgium.
I was then working at the Medical Protein Chemistry lab of Prof. Joel Vandekerckhove.

The aim of this applet is simple: to get a more accurate mass-measurement from a MALDI-TOF mass-spectrometer.

In the lab we noticed that the linear calibration (two calibration masses) yielded the largest inaccuracies in between the two masses. Since accurate measurements are essential to protein identification using mass-spectrometry (increasingly so as genetic and protein sequence databases grow!) this inaccuracy was rather annoying.

The idea was that a poly-alanine spectrum covered the mass-range more closely (instead of two calibration points, you'd have about thirty!) so it would yield better mass-measurements when used for calibration.

Unfortunately the software on the MALDI-TOF we used did not allow us to calibrate using poly-alanine.
The workaround was then to use standard linear calibration to record a poly-alanine spectrum, and then calculating the error on each of the poly-alanine peaks, plotting these versus the masses of the peaks and using the resultant 'error to mass' plot (red dots in the graph) to construct a regression curve (blue curve in the graph) that would predict the error for a certain mass.
The sample to be identified is then recorded with the same linear calibration as the poly-alanine spectrum.
The final step consists of subtracting the predicted error for a certain peak from the measured mass. The result turns out to be up to twice as accurate.

An added feature in this applet is the possibility to use any sequence of amino acids as calibrant. In the textfield "sequence of calibrant" you can put the one-letter amino acid code of your calibrant of choice. The conventions for doing this are discussed next.

Conventions for using your own calibrant
You can enter a single amino acid or multiple amino acids for each entity described as 'sequence block'.
The maximum number of sequence blocks is three.
The minimum is one.
Sequence blocks are separated by spaces. Inside a sequence block no spaces or other separators are allowed.

There are three possible ways to enter sequence blocks:
Note: the examples in the following list are representations of the actual sequence, NOT of the desired input. Examples for input are provided at the end of the following list.

• One sequence block that is repeated (e.g.: AAAAAA...; ARNDARNDARND...).



• Two sequence blocks: the first is constant, the second is repetitive (e.g.: R AAAAAA...; ARND GHIKKGHIKKGHIKK...). Please note that the constant part can be a C-terminal or N-terminal subsequence, yet for our calculation that doesn't matter (only the mass does) so be careful to always enter the constant part first.



• The third one is for ease of use only, since it is equivalent to the second way: it states a constant block, a repetitive block and finally a second constant block (e.g.: R AAAAAA... K; ARNDG KLMKLMKLM... HHGRA).
  Of course this is equivalent to 'RK AAAAA...' and 'ARNDGHHGRA KLMKLMKLM...' as described in way two, yet it may be easier to read.

Since the MALDI-TOF we used showed an error distribution that resembled a cubic curve, we used cubic regression analysis, so this is the standard setting. You can alter this by changing the number in the textfield on the right. Pressing 'Calibrate' again will recalibrate using the new degree of regression.

A test spectrum of Arg-poly(Ala) (use default R A setting as calibrant sequence) is available as testCalib.txt or testCalib.xls from http://home.tiscalinet.be/LennartMartens/resources/. Just cut and paste the contents of this file in the calibration window and run.
You can also do the correction on the same spectrum. Just paste in the peakfiles window and correct after calibrating.

Finally, any questions, comments, suggestions etc. are welcome at: