Welcome to AP-iNET

AP-iNET is a computational prediction system for host-pathogen protein-protein interactions in the Arabidopsis-Pseudomonas model system. This webtool allows users to input host and pathogen protein sequences to predict whether they are interacting within a host-pathogen network.

Protein-protein interactions (PPIs) play a crucial role in host-pathogen interactions and pathogenesis and maintaining infection. Successful prediction of PPIs for a well-studied host-pathogen system will lead to better understanding of mechanisms of pathogenesis on a genome-wide scale.

Several support vector machine(SVM) models is developed using experimentally proved Arabidopsis – Pseudomonas PPIs from different well curated host-pathogen interaction (HPI) databases.These models are based on Amino acid composition (AAC), Dipeptide composition (DIPEP), Conjoint Triad composition (CT) and Composition-Transtion-Distribuition (CTD) features. The SVM models are validated through a five-fold traing dataset and independent dataset. User has a choice to use individual descriptors or in a combination for prediction of the host-pathogen interactions. A homology based model also developed to assess the predictions in Arabidopsis - Pseudomonas system in genome wide. More information about these models can be found on the Help page.

There are three options for interaction prediction: “Host Only”, “Pathogen Only” or “Host and Pathogen”. User can submit the protein sequences in FASTA format by either uploading file or through a text box. After the prediction model selected and sequences uploaded, the submitted sequences are tested against trained files of the selected model. If both host and pathogen sequences are inputted, all possible combination of host-pathogen sequence pairs will be created and tested against the trained SVM model. After the prediction is complete, the prediction results will be displayed in a table of all posible pairs with their svm scores and the reamrk of interacting or non-interacting. The output can also be downloaded in various fourmats (CSV,XML,TSV) and the network of interaction can be viewed through "View Cytoscape".