This procedure is used to compute ratios of peptide and protein abundances and then to perform a statistical analysis. The significance of ratios varations is then computed at protein and protein levels using two statistical tests (T-Test and Z-Test). Several filters can also be set to increase the quality of quantitative results.

Here is the description of the parameters that could be modified by the user.

* Use only specific peptides: if checked, peptides shared between different protein sets will be discarded from the statistical analysis.

* Discard missed cleaved peptides: if checked, peptides containing missed cleavages will be discarded from the statistical analysis. It has to be noted that perfect tryptic peptides whose sequence is included in an observed missed cleaved peptide are also discarded if this option is enabled.

* Discard oxidized peptides: if checked, peptides containing the Oxidation(M) modification will be discarded from the statistical analysis. It has to be noted that non-modified peptides whose sequence is the same than an observed oxidized peptide are also discarded if this option is enabled.

• Normalisation:

par run en prenant une référence (1er run); sur la médiane de la distribution des ratios calculés entre chaque run et la référence

• Statistical tests alpha: the p-value threshold used for the two statistical tests

• T-Test: a Welch's modified T-Test (allows for unequal variance) used to assess the significance of the abundance mean difference between compared sample groups.

• Z-Test: Z-scores are computed through a standardization procedure of observed log ratios ⇒ (log ratio - median_ratio)/log_ratios_standard_deviation. Z-scores (or standard scores) corresponds a number of standard deviations in the gaussian standard distribution. They can thus be easilly converted into p-values.

Peptide abundances can be summarized into protein abundances using several mathematical methods:

• sum: for each quantitative channel (raw file) the sum of observed peptides abundances is computed

• mean: for each quantitative channel (raw file) the mean of observed peptides abundances is computed

• mean of TOP3 peptides: same procedure but applied on the 3 most abundant peptides. Peptides are sorted by descending median abundances (computed accross all compared samples for peptide). Then the 3 first peptides are kept.

• median: for each quantitative channel, the median of observed peptides abundances is computed

• median profile: a matrix of peptide abundance ratios is first computed (rows correspond too peptides and columns to quantitative channels). The median of these ratios is then computed for each column. The relative values are then converted back into absolute values using a scaling factor. This factor is computed as the maximum value from the means of TOP3 peptides abundances.

• normalized median profile: matrix of peptide abundance ratios is first computed (rows correspond too peptides and columns to quantitative channels). This matrix is then normalized and then summarized using the median method described above. The obtained median abundances are then adjusted by using a scaling factor. This factor is computed as the maximum value from the means of TOP3 peptides abundances.