GO Browser
==========

Provides access to Gene Ontology database.

**Inputs**
- Cluster Data: Data on clustered genes.
- Reference Data: Data with genes for the reference set (optional).

**Outputs**
- Data on Selected Genes: Data on genes from the selected GO node.
- Enrichment Report: Data on GO enrichment analysis.


**GO Browser** widget provides access to [Gene Ontology database](http://geneontology.org/). Gene Ontology (GO) classifies genes and gene products to terms organized in a graph structure called an ontology. The widget takes any data on genes as an input (it is best to input statistically significant genes, for example from the output of the **Differential Expression** widget) and shows a ranked list of GO terms with p-values. This is a great tool for finding biological processes that are over- or under-represented in a particular gene set. The user can filter input data by selecting terms in a list.

![](images/go_browser/GO-Browser-stamped.png)

**INPUT tab**

1.  Information on the input data set. *Ontology/Annotation Info*
    reports the current status of the GO database.
2.  Select the reference. You can either have the *entire genome* as
    reference or a *reference set* from the input.
3.  Select the ontology where you want to calculate the enrichment.
    There are three *Aspect* options:
    -   [Biological
        process](http://geneontology.org/page/biological-process-ontology-guidelines)
    -   [Cellular
        component](http://geneontology.org/page/cellular-component-ontology-guidelines))
    -   [Molecular
        function](http://geneontology.org/page/molecular-function-ontology-guidelines)

4.  A ranked tree (upper pane) and list (lower pane) of GO terms for the
    selected aspect:
    -   **GO term**
    -   **Cluster**: number of genes from the input that are also
        annotated to a particular GO term (and its proportion in all the
        genes from that term).
    -   **Reference**: number of genes that are annotated to a
        particular GO term (and its proportion in the entire genome).
    -   **P-value**: probability of seeing as many or more genes at
        random. The closer the p-value is to zero, the more significant
        a particular GO term is. Value is written in [e
        notation](https://en.wikipedia.org/wiki/Scientific_notation#E_notation)).
    -   **FDR**: [false discovery
        rate](https://en.wikipedia.org/wiki/False_discovery_rate) - a
        multiple testing correction that means a proportion of false
        discoveries among all discoveries up to that FDR value.
    -   **Genes**: genes in a biological process.
    -   [Enrichment](http://geneontology.org/page/go-enrichment-analysis)
        level

![](images/go_browser/Filter-Select.png)

**FILTER tab**

1.  *Filter GO Term Nodes* by:
    -   **Genes** is a minimal number of genes mapped to a term
    -   **P-value** is a max term p-value
    -   **FDR**: is a max term [false discovery
        rate](https://en.wikipedia.org/wiki/False_discovery_rate)

2.  *Significance test* specifies distribution to use for null
    hypothesis:
    -   [Binomial](https://en.wikipedia.org/wiki/Binomial_distribution):
        use a binomial distribution
    -   [Hypergeometric](https://en.wikipedia.org/wiki/Hypergeometric_distribution):
        use a hypergeometric distribution

3.  [Evidence codes in
    annotation](http://geneontology.org/page/guide-go-evidence-codes)
    show how the annotation to a particular term is supported.

**SELECT tab**

4.  *Annotated genes* outputs genes that are:
    -   **Directly or Indirectly** annotated (direct and inherited
        annotations)
    -   **Directly** annotated (inherited annotations won't be in the
        output)

5.  *Output*:
    -   **All selected genes**: outputs genes annotated to all selected
        GO terms
    -   **Term-specific genes**: outputs genes that appear in only one
        of selected GO terms
    -   **Common term genes**: outputs genes common to all selected GO
        terms
    -   **Add GO Term as class**: adds GO terms as class attribute

Example
-------

In the example below we have used **GEO Data Sets** widget, in which we
have selected *Caffeine effects: time course and dose response* data set, and connected it to a **Differential Expression**. Differential
analysis allows us to select genes with the highest statistical
relevance (we used ANOVA scoring and agent label) and feed them to **GO Browser**. This
widget lists four biological processes for our selected genes. Say we
are interested in finding out more about *monosaccharide transmembrane transport* as
this term has a high enrichment rate. To learn more about which genes
are annotated to this GO term, select it in the view and observe the results in a **Data Table**, where we
see all the genes participating in this process listed. The other output of **GO Browser** widget is enrichment report, which we observe in the second **Data Table**.

![](images/go_browser/GO-Browser-Example.png)