Tutorial

1 Getting started: single tissue query


1.1 To use FNTM, first select a tissue and input one or more genes.


1.2 Mammary gland has been selected along with Brca1 and Brca2 here, and should look like this:


The network displayed indicates the most functionally related genes to Brca1 and Brca2 in the mammary gland. Slide the 'Minimum relationship confidence' and 'Maximum number of genes' sliders to adjust how many genes (nodes) and functional relationships (edges) are shown.

1.3 Hover over an edge to find out the relationship confidence and what datasets contribute to the functional relationship prediction for a displayed edge between genes:


The relationship confidence between Brca1 and Exo1 is close to 1 (0.98), which indicates that the prediction is very confident. At the top of the table, horizontal bars represent the proportional contribution of different types of datasets. In this example, the predicted relationship between Brca1 and Exo1 is mostly based on co-expression in microarray datasets (80.9%), but these genes also have evidence of functional relationship in three other datasets. The top individual dataset contributions are also displayed, here showing that the non-expression datasets are individually the strongest evidence for the edge, while the large number of expression datasets with relatively weaker evidence in each dataset contribute most of the evidence when taken together.


1.4 Under the displayed network, the table shows the most enriched terms in the Gene Ontology (Molecular Process branch) and KEGG pathways:


The enriched terms help to interpret the network of query and related genes. In this case, the most enriched terms reflect the role of Brca1 and Brca2 in DNA repair, and it's unsurprising that many of the related genes would also be involved in DNA repair and other DNA replication processes. The related genes that are shown in the network but not annotated to an enriched GO term could be candidate genes for additional experiments to see if they are involved in these processes or are functionally related in some other way.

1.5 Additional information about the query and related genes is avaliable to the right of the displayed network:


At the top, the search can be modified to include different tissues or genes. Next, the current query input genes are listed with links to other gene databases and a check mark to indicate which genes are annotated to the tissue according to MGD-GXD. All of the related genes are listed in the "Related Genes" table, with columns indicating the average relationship confidence to the query genes and whether the gene is annotated to the tissue according to MGD-GXD. The "Relationships" table lists all of the displayed relationships and relationship confidence with the query gene set.


2 Translational example: Studying the role of Ptgs2 (Cox2) in two different tissue-specific phenotypes.


2.1 In this example, a translational researcher is interested in the role of Ptgs2 in epilepsy, and would like to identify candidate genes to study that may be functionally related to Ptgs2 specifically in the brain. As Ptgs2 is also known to promote angiogenesis, the researcher queries for Ptgs2 in both the brain and blood vessel tissue networks here:


The network displayed indicates the most functionally related genes to Ptgs2 in the brain and blood vessel. Genes that are displayed in both networks are colored orange and move in tandem to keep their positions in both networks coordinated. The 'Minimum relationship confidence' and 'Maximum number of genes' sliders can be adjusted independently for each network to change how many genes (nodes) and functional relationships (edges) are shown.

2.2 Several visualization controls for multi-tissue view are on the left side of the page:


As in single-tissue view, the current search can be modified to include a different combination of tissues and genes. The query genes are listed along with their short descriptions. Individual tissue networks can be toggled on and off to make it easier to compare tissues if many tissues have been queried at once. The gene set enrichment tables as well as related genes and relationships to query genes are displayed if the researcher checks the "show gene tables" checkbox.


2.3 Checking the "show gene tables" checkbox reveals the gene set enrichment tables for the query and related genes in each tissue:


Even though the researcher has not specified the processes they are interested in, the tissue-specific networks reflect that Ptgs2 has different roles in different tissues by predicting functional relationships between Ptgs2 and many angiogenesis genes in the blood vessel network, and many inflammatory response genes in the brain network. If the researcher is specifically interested in epilepsy, they may want to screen the other related genes in the brain network and use the related genes in the blood vessel network as a control, possibly in combination with a Ptgs2 inhibitor.