Principle

---

Peptide mass fingerprinting involves the digestion of an unknown protein with a proteinase of known cleavage specificity and the measurement of the resulting peptides by mass spectrometry. Those experimentally measured peptide masses are then compared with the theoretical peptides calculated for all proteins in a protein sequence database.

One of the main ideas of Aldente is to avoid calibration problems by taking into account the mass spectrometer deviation. It implements a number of rules, empirical observations and user knowledge that approach the expert human interpretation of results in various steps of the identification procedure. The procedure can be divided into several steps:

• The selection of experimental MS peaks to be matched with theoretical peptide masses within a user-defined tolerance space.
• The interpretation of the experimental calibration deviation of the instrument: Exclusion of false matches and reduction of the tolerance space to the internal deviation of the instrument.

Aldente proposes a realignment of experimental masses using the Hough tranform. The Hough Transform is a standard method used in image analysis for finding straight lines hidden in larger amounts of other data. It is a robust method, therefore not sensitive to noise. It has no difficulties to work with very crowded spectra (i.e. with more than 100 input masses) and with a lot of theoretical masses (that can be generated when considering the combinatory related with heterogeneously modified or missed-cleaved peptides).

The output provides a significant amount of useful information. It includes identified proteins and peptides, as well as the deviation function used, along with the ambiguities that have been worked out.

Precision and Accuracy

---

Most of the time, mass spectrometers have very good precision but poor accuracy producing measures with calibration errors.

Recalibration

---

A simple calibration using a known sample is not possible, because the accuracy changes from one shot to another. A solution should be to add in the sample known peptides and use them to recalibrate the spectrum, but this is not always possible.

The calibration errors or deviation function of the mass spectrometer can be approximated by a line: measured masses = ( slope x real masses ) + shift Aldente estimates this function from the data and it is able to work with non recalibrated spectra.
Aldente compares each protein from the database against the experimental data. Here is one example of all possible matches between theoretical and experimental masses.
The user defines the maximum delta allowed to compare the masses. The user should have an idea of the maximum calibration errors of the machine and be able to limit the alignment with allowed slope and shift range. Shift range is defined in Daltons, Slope range is defined in ppm.
Aldente locates and solves the ambiguities.
Aldente optimizes the set of best matches. The hypothesis is that the mass spectrometer deviation is the line passing by the maximum of plots. Using Hough transformation, the tool will find this alignment and define the good matches while excluding the outliers.
This representation scales the axis differently, so that the internal error of the mass spectrometer can be represented. The user defines the mass spectrometer internal error in ppm, and it is taken into account in the alignment.