Welcome to MAP
MAP (Model-based Analysis of Proteomic data), is designed to statistically compare the proteomic profiles generated from different biological samples using the isotope labeling based mass spectrometry (MS) technique and directly identify proteins with significant abundance changes. Unlike many existing tools for this purpose, it does not require parallel/additional technical replicates to fathom technical variations; instead, MAP uses a novel step-by-step regression analysis to directly model technical variations from the profiles under comparison. Therefore, experimental designs and their expenses can be simplified and reduced for more practices.
Basic modules of web application1. MAP case-control Module
It takes two proteomic profiles generated from different samples as input, in which the MS intensities could be assigned to either protein level or peptide level. It will report a table of statistics, which include the log2-ratio of normalized MS intensities for each protein/peptide and a P-value to describe the significance of its intensity change as well as adjusted P-value (by BH correction), together with three figures describing the whole analysis.2. MAP integration Module
In this module, MAP can compare the proteomic profiling data of two different samples with multiple biological replicates (typically generated in different MS runs). Before running this module, the user first needs to repeatedly perform pairwise cross-sample (case-control) comparison with the profiles generated in the same MS run using the first module, and collect the output table of each comparison. Then, the integration module takes these tables as input and reports the kinds of statistics accompanying with their P-value to evaluate the difference of each protein between two conditions. Here, we recommend to use the second best P-value to rank and select the final differentially expressed proteins if the number of MS runs not exceeding three, otherwise, there being many MS runs, for example, 4, 5, ..., we recommend to apply average Z-statistic to screen the final results.
Currently, this work is under review. One of the projects was applied MAP (some of them used the preliminary version of it) has been published.