Hemagglutinin modeling server


Flu3D is a web-service designed to generate useful information about Hemagglutinin (HA) proteins for influenza virus type A (IVA).


The user can submit a new request by clicking the Send new request button on the web-server homepage or by clicking on the Submit request field in the upper bar. In the submission form, the user can upload a FASTA file containing either single or multiple HA sequences. The user should also specify the name and the description of the request. After uploading the file the request is sent clicking on the button. The user can run a sample request by clicking on the button or visualize an example of the results clicking on the See example results button. Both buttons are present in the request page. After submission, the request is added to the queue and a link to the request status page page is given. This page allows to check the status of the request, provides the link to the visualization page and allows to download the results of the computation. The latter two are displayed only after the successful completion of the calculations.
The server performs quality control on each submitted sequence and selects the closest HA template to such sequence (using BLAST). The server performs homology modeling with the help of Modeller. The electrostatic potential (ESP) of each generated model is then computed using APBS. The sequence, model and ESP information previously obtained are then used to compute the distances between each query and all the protein structures already collected.


The distances previously computed (sequence distances, structure RMSD and ESP distances respectively) are further used for the dynamic graph representation of the proteins. Each node in the graph represent a protein. The vicinity of two nodes is inversely proportional to the distances previously computed (e.g. RMSD). After the completion of the calculations the user can visualize the position of its proteins with respect to other ones clicking on the visualization link appositely generated. The user can also explore the proteins stored in the database directly, by selecting the Basic Visualization field in the upper bar. Such visualization doesn't require any request. The protein structures can be visualized clicking on the respective nodes in the graph. When multiple structures are selected, the multiple sequence alignment of their respective sequences is shown in the panel below the structural representation. Metadata information (e.g. protein name, IVA strain, year and location of infection, etc.) can be visualized passing the cursor over the nodes. Both the menus and the search bar improve a friendly exploration of the protein network. For instance groups of proteins sharing some properties can be easily highlighted using the tools previously mentioned.