About

FunGeneNet is a web tool aimed at proving the functionality of networks in a wide range of sizes of experimental gene sets, both for different global networks and for different types of interactions. It utilizes the ANDSystem and STRING to reconstruct gene networks using experimental gene sets, provided by the user, and to estimate their difference from random networks. A main feature of the method implemented in FunGeneNet is that it allows consideration of specific types of molecular-genetics interactions, while most of the existing methods only work with generalized types characterizing the existence of any interaction between network members. The sensitivity of our method exceeds 0.8, the specificity is 0.95.

Help

1. On the first step a user need to select which base of knowledge (STRING or ANDSystem) should be used as a source for the networks generation.


By selecting STRING as a source of networks, all vertices will be linked by a generalized type of interaction, while in the ANDSystem they can be linked by various types of interactions.
The main menu of FunGeneNet

2. Depending the selected knowledge base a menu will be expanded with corresponding data.


To perform an analysis the user need to upload his experimental set of genes/proteins, by pressing the 'UPLOAD' button and providing ".txt" file, an example of such file for ANDSystem is available here, while a file for STRING can be downloaded here. Also these examples can be obtained by pressing the "EXAMPLE" button in the menu.

2a. ANDSystem


In case of ANDSystem the FunGeneNet allows to perform analysis of uploaded experimental gene/protein sets, corresponding to the one of species from the "ORGANISM" dropdown list.

Next, the user need to specify which database his list belongs to in the "DATABASE" dropdown list.

Please note, that when "DATABASE" list is opened, you can use a mouse trackball for the navigation through the rest of external databases presented inside this list.


Also, when ANDSystem is selected, the user can select a type of interactions inside the "INTERACTION TYPE" list.

There are two operation modes in FunGeneNet, when ANDSystem is selected. The frst mode is applied when "all types" for the interaction is selected. In this case, all interactions are considered as a generalized type of interaction. The second mode is applied when a specified type of interaction is selected (for example, "activity regulation and transport", "catalysis", etc.). In this case, the system takes into account only interactions of the specified type, while any others are removed from the network.

An example of FunGeneNet menu when ANDSystem is selected

2b. STRING


In case of STRING the current version of FunGeneNet allows to perform analysis of uploaded data corresponding only to the "Human", thus, the first dropdown list contains only one organism, selected by default.

Similar to the case when ANDSystem is selected, the user need to specify which external database his list belongs to in the "DATABASE" dropdown list.


The third dropdown list allows user to select a score value (400 by default), which describes the weight of interactions and allows to eliminate noise information.

Please note, that increasing the score value will result in reducing the number of false interactions, but can also decrease the completeness of networks.

An example of FunGeneNet menu when ANDSystem is selected

After all required field are selected and the TXT file with a list of analysed gene/protein IDs is uploaded to the server a "CONTINUE" button will become active, by pressing it an analysis will starts, the user will be forwarded to a new page with the analysis progress.


3. Analysis


The page contains two tables the first one indicates an analysis progress, including which step is currently performed by FunGeneNet and how many steps left, while a second one provides user with analysis results, containing information about t-test p-value (this value characterizes the difference between the analysed network and generated random networks), ROC curve p-value (a proportion of negative sample networks having a t-test p-value less than or equal to that for the network being analyzed) and a file containing with ANDSystem or STRING network (regarding the selected knowledge base), corresponding to the uploaded data and selected interaction type.

Pressing the "CONTINUE" button, will visualise the reconstructed network.

An example of FunGeneNet analysis using ANDSystem