Understanding and modeling of cellular processes depend on comprehensive information of protein networks. Large-scale affinity purification coupled with mass spectrometry (AP-MS) provided comprehensive data for the analysis of protein complexes. In large-scale AP-MS experiment, there are many different conditions in which different proteins are tagged, and in each pull-down there is high number of proteins which include a lot of contaminants. So dealing with this large amount of data to infer a reliable protein-protein interaction network is an essential task. Here, we propose a new algorithm which uses the concept of information theory for analyzing the parallel proteomic data. Information-theoretic methods use mutual information, which is an information-theoretic measure of dependency. Mutual information is being used for calculating the association score of each protein interaction based on measuring the similarity of protein profiles among different pull-downs. So with this algorithm we will be able to infer protein-protein interaction network with weighted edges using quantitative mass spectrometry, in which the weight of each interaction indicate the probability of the occurrence of that interaction.