Functional Analysis and Visualization for large scale Omics data
FuncTree is a web based application, which allows user to visualize, customize, and compute statistical test to understand the biological functionality of their omics data.
FuncTree allows user to map their omics data on to a pre-defined treemap, which is based on the KEGG brite functional hierarchy.
This allows user to quickly and comprehensively understand the functional potential of their data, and to develop further hypothesis and scientific insights.
Each node in the treemap represents a particular biological function, which is defined by the KEGG database.
Edges between nodes represents the hierarchical relationship between functional units.
This visualization method allows users to graphically visualize their data's functionality across multiple functional layers, which enables users to overview and comprehend the broad functional potential of their data.
FuncTree's treemap combines the KEGG brite functional hierarchy database, KEGG pathway database, KEGG module databasee, and KEGG orthology database.
Each layer of the circular dendrogram represents the different functional layer of the KEGG functional hierarchy.
FuncTree user interface
The main display displays the "Functional Tree" map representing KEGG brite functional hierarchy as a circular dendrogram. Navigate the Treemap using basic pan/zoom controls.
Mouseover on each node to display a tooltip showing the name of the biological function it represents.
To get detailed information of the node, click "Node Information" on the menubar, described below (6).
Click on each node to collapse/expand subordinate nodes under it.
Click on the selection tab to swith between the Treemap and Expression Table. Table will upload dynamically as you map your data on the Treemap.
Upload data via Text query or File upload. Click the "Map input data" button to map data on the Treemap.
See "Input Data Type" section for detailed informations.
FuncTree provides extensive customization options to achive optimal visualization for various data. Clicking on the "Option" tab on the left menu bar will open up the option panel, which provides access to all available customization options.
To aply customization, simply check the checkbox of the option that you want to aply. Some options allows even more detailed customization to achieve optimal result.
See "Map Customization" section for detailed information about each option.
Click the "Update option" button to reflect the customization.
Customized Treemap in FuncTree can be exported in SVG format.
Expression Table can also be downloaded in txt format.
Clicking on the "Download" tab on the left menu bar provides access to the download panel. Select the appropriate data format to download the file.
Further downloadable content and selectable file format is being developed for future version update.
Expression Table
Clicking on the "Node Information" tab, enables you to explore further detailed information about each node.
- Name: The name/id used in KEGG database.
- ID: Internal ID for the FuncTree treemap.
- Definition: Definition of the node's biological function.
- Input: Value that user specified directly. 0 if no specification.
- Calculated: Value that is calculated using specified calculation method.
- Mapped: Finalized value that is mapped. This can be equivalent to "Input" "Calculated" value depending on your option. (0 ~ 200)
- Opacity: 0 (full transparent) ~ 1 (fully opaque)
- Color: Color of the node in hexadecimal. (#ff0000)
- Node Description: Detailed description of the node's biological function.
The general idea of FuncTree's mapping methodology is
- Select a particular node (or biological function) by specifying the node's name/id.
- Specify a particular value (size of the circular mapping) to the node. We recommend using value between 0 ~ 200 for optimal visualization.
- (Optional) Specify the opacity value of the node. Value must be in the range between 0 ~ 1.
- (Optional) Specify the color of the node in hexadecimal (#ff0000).
FuncTree supports the following data types on the right to customize the treemap. Make sure you use the required prefix for each data type, as shown in the examples below.
- Text input
- File input
FuncTree enables user to directly customize the treemap using a simple text query. Each attribute must be separated by a "space". See examples below.
| Input | Description |
|---|---|
| n-map00010 v-120 | map circle of size 120 on node "map00010" |
| n-1881 v-78 o-0.3 n-map00030 v-12 |
map circle of size 78 with opacity value 0.3 on node "1881" ("M00281") map circle of size 120 with default opacity value(1) on node "map00030" |
| n-map00190 v-149 o-0.9 c-#ff0000 | map red circle of size 149 with opacity value 0.9 on node "map00149" |
FuncTree also allows user to upload large dataset via file upload. The file must be in TSV format in order to be visualized correctly. Download example file below.
| Data Type | Prefix | Example | |
|---|---|---|---|
| Node | ID | n- | n-7322 |
| Name | n- | n-map00230 | |
| Value | v- | v-150 | |
| Opacity (optional) | o- | o-0.6 | |
| Color (optional) | c- | c-#ff0000 | |
- Pipeline In order to visualize large scale omics data, FuncTree expects input data to be processed along the following pipeline.
- Aquire short read DNA sequence using high throughput DNA sequencing technology.
- Annotate each short read sequence into KO (KEGG Orthology) using annotation server such as KAAS.
- Normalize the value of each KO according to the number of read annotated to it. This will create a list of KO and it's relative abandance.
After you have created the KO list, choose one of the following options. - Calculate the value of functions in upper layer (Module/Pathway) using own calculation method. After you have created a list/table of all nodes with designated attributes (value, opacity, color), directly map the result using FuncTree via file upload.
- Calculate the value of functions in upper layer (Module/Pathway) using FuncTree's calculation method by checking the "Internal node calculation" option in the Option menu.
- Calculate the statistical significance of your data by mapping your file and checking the "Enrichment analysis: Metagenome" or "Enrichment analysis: Genome" option. Value of internal nodes (Module/Pathway) would be calculated automatically.
- Direct Mapping
- Calculate the relative abundance of each KO from short read DNA sequence, using server such as KAAS.
- Calculate the relative abundance of functions in upper layer (Module/Pathway) using own calculation method.
- (Optional) Normalize the relative abundance of each function, so that the value will be in range between 0 ~ 200.
- (Optional) Assiciate each function with a color and an opacity value.
- Map the list/table on to FuncTree via file upload. Do not forget a prefix before each attribute.
- Node Calculation
- Statistical Calculation
Click here to download example file.
FuncTree can also provides its own calculation method to automatically calculates the value of internal nodes (Module/Pathway). Upload a KO list and select a calculation method to quickly reconstruct and visualize the functionality of all of the functional layer.
FuncTree allows user to calculate the statistical significance of their omics data. This is useful when you want to identify unique characteristic of your omics data. FuncTree uses Fisher's exact test to calculate the p-value of each node. By default FuncTree allows user to statistically compare their metagenomic data with metagenomic data aquired in the Human Microbiome Project.
In order to conduct statistical calculation, simply upload your KO list via file upload, and in the Option menu check the Enrichment analysis checkbox (Metagenome/Genome depending on your data). Select a background data that you want to compare your data with, and select a statistical correction method for multiple testing. You may also set a coverage value for Module and Pathway to filter out any node which has low coverage.
If you do not want to use FuncTree's calculation method or if you want to compare your data with other background dataset, we recommend you to calculate the p-value of each node using own calculation method and mapping the result using Direct Mapping.
Direct Mapping of human gut metagenomic data
Functions unique to human gut metagenome, mapped using statistical comparison with metagenomic data from other body sites.
FuncTree provides extensive customization options to achive optimal visualization for various data. Clicking on the "Option" tab on the left menu bar will open up the option panel, which provides access to all available customization options.
- Input value normalization
- Visual normalization
- Undefined KO (Default)
- Scale
- Linear (Default)
- Sqrt
- Domain range
- Internal node calculation
- Internal node calculation method
- Average (Default)
- Summation
- Average/Summation * Node coverage
- Node value prioritization
- Set coverage
- Enrichment analysis: Metagenome/Genome
- Background data
- Multiple testing method
- Coverage
- Recenter functional tree
Normalize the input data by scaling each value into alignment. This allows FuncTree to visualize proportionally correct treemap result, which allows the comparison of
treemap for different datasets.
Normalize the input/calculated value of each node so that for each functional layer, value will be in the range of 0 ~ 200.
Currently there are 7547 KOs which are not categorized to any Module/Pathway. FuncTree allows user to choose whether to include these undefined KOs into account when normalizing the node's value. Having this option turned on, could result in a treemap which is statistically correct but visually incomprehensible, as a consequence of these undefined KO having large value. Checking this option could result in a treemap which could be statistically misleading but visually comprehensible.
Scale the size of circle so that the "radius" of the circle would be proportionate to the value.
Scale the size of circle so that the "area" of the circle would be proportionate to the value.
It may be the case that you only want to visualize functions within a specific range. Check the Scale option (Linear/Sqrt with limited domain) and specify a domain range to aply scale (Linear/Sqrt) within a specific range.
Ex. Linear scale with limited domain + Domain range (20,80)%
Aply Linear scaling to nodes that have a value within the bottom 20% and top 80% of that functional layer.
Calculate the value of internal node (Module/Pathway) from KO list.
Value of upper nodes (Module/Pathway) will be the average value of the KO assigned to that node. Average is defined as the total sum value of the KO assigned to that node devided by the total number of KO with value (node value≠0).
Value of upper nodes (Module/Pathway) will be the summation value of the KO assigned to that node.
Value of upper nodes (Module/Pathway) will be the product of average/summation and coverage. Coverage is defined as the number of KO with value (node value≠0) devided by the total number of KO assigned to that node.
Ex. If average KO value under node "map00010" is 120 and 10 out of 45 KOs have value, the value of "map00010" will be 120 × 10/45 = 26.7
If a node has both a value which was directly attributed by the user and also a calculated value based on the KO value assigned to that node, this option sets which value the user would like to prioritize. The default is set to prioritize user attributed "Input" value.
Only visualize nodes above the designated coverage value. Default visualizes nodes with coverage value ≧ 0 (all nodes). Coverage is defined as the number of KO with value (node value≠0) devided by the total number of KO assigned to that node.
Identify statistically significant functions in your data by conducting enrichment analysis. Input data must be a KO list of metagenomic/genomic dataset.
Select a background data that you want to compare your data against.
Select a statistical correction method for multiple testing.
Only visualize nodes above the designated coverage value. Default visualizes nodes with coverage value ≧ 0 (all nodes). Coverage is defined as the number of KO with value (node value≠0) devided by the total number of KO assigned to that node.
Redraw the treemap with the designated node at the center of the treemap. This allows user to draw treemap which focuses on a particular category.
Treemap fucusing on Metabolism
Treemap fucusing on Genetic Information Processing
FuncTree is supported by the following web browser. We recommend using Google Chrome or Safari for best performance. Certain visualization may not be rendered properly depending on the version of your browser.
- Windows: Google Chrome, Firefox, Opera
- Mac: Google Chrome, Firefox, Opera, Safari