Thank you for your prompt response.
For the comparison of two conditions does the program use the paired t-test? Another question is if the t-test is assuming technical replicates and biological replicates or if it uses all the values all together indepently of being technical or biological replicates?
I do not know if I made myself clear, but when we do the t-test in excel we obtain different p-values, and this is why we have some doubts.
Thank you. |
If you have specified something for the "Identity Annotation" in the group comparison, then the replicates get grouped together based on that annotation, and the values for each get averaged together before being used in the t-test.
If you are calculating the fold change "per peptide" (and not "per protein"), then you can use the document grid to see the normalized area values that are used in the t-test. Just make sure that the normalization method on:
Settings > Peptide Settings > Quantification
is the same as the normalization method specified for your group comparison.
In the Document Grid, you can customize the view and find the "Normalized Area" column (there's a search button on the toolbar in the View Editor-- it looks like binoculars). The "Normalized Area" column is under "Peptide Results/Quantification".
In order to calculate the fold change, Skyline takes the logarithm (base 2) of the normalized areas for each of the replicates. If there are technical replicates, then those log areas are averaged together for each group of technical replicates.
Then, a linear regression is performed where the x-value is 0 or 1 depending on whether the replicate was in the control group, and the y-value is the log area. Skyline raises 2 to the power of that slope, and that is the fold change for that peptide.
No, it is not a paired t-test. |