| EcoPPI aims to provide
a computationally derived protein-protein functional linkages in Escherichia coli.
It allows users to search the protein synonyms to identify its functionally linked proteins using
Go to SearchDB tab. Fill up the gene name or orf name in query text box on the page. Then hit submit button to search the database. Search result The EcoPPI will return the the result of Database search back to the browser as table in HTML page. Each entry in result table is a interacting partner of query gene or protein. PCC is a value of gene expression correlation coefficient between query gene or protein and interacting partner. PCC is calculated using gene expression profile in 380 conditions of a pair of genes The side panel of EcoPPI page is linked to our institute's website and also to other web tools developed by our institute. How EcoPPI was generated? Methods:Six protein-protein functional linkage prediction methods used to generate confidence scores for all possible pairs of Escherichia coli proteins which includes phylogenetic profile, gene cluster, minimum gene distance, gene order conservation, expression profile similarity and improved mirrortree methods. These confidence scores used to train seven machine learning classifiers on gold standard dataset. These seven classifiers are Support Vector Machine, Naive Bayes, Bayesian Network, Decision trees, Random Forest, Logistic Regression and Neural Network. Then for each classifier we predict genome-wide protein protein interactions and interactions predicted by four or more classifiers are considered as true interacting pairs. Due to a combination of six different interactions prediction methods and a combination of seven classifiers, our predicted dataset is highly relible dataset among other prediction methods After prediction, we have done following analyses-
|