Data submission
Results
Network visualization
Candidate mechanism retrieval
Data submission
The image below shows the first impression that you get from NeuroMMSig. In the navigation bar clicking in
"Quick View" , one can get access the "Introduction" and "How to use" sections plus
database login for developers. Below, the NeuroMMSig form where one should select at least one disease
context and introduce the data. Select multiple disease allows you to compare subgraphs between them. Thus,
one can compare subgraphs from Alzheimer's and Parkinson's disease. Multiple data types can be sent by
clicking in the "+" button and adding multiple inputs to the form.
The "settings" button enables you to optimize the algorithm parameters. By default all variables
are set to 1. In order to submit data you have to press the enter button or click on the submission
button. The last button enables you to change the context of the query using the "OR" and
"AND" options. The "OR" options can be used to retrieve the union of all subgraphs in the
query and the "AND" to retrieve the intersection of all subgraphs in the query (only the subgraphs
that match all data will be returned, similarly to a MySQL query).
In the table below, all types of data accepted by NeuroMMSig together with their format are listed. Once data
have been introduced, please press ENTER or click the "Submit" button
Data type |
Format |
Example |
Gene |
HGNC Symbols comma separated. HGNC is responsible for approving unique symbols and names for human
loci,
including
protein coding genes, RNA genes and pseudogenes, to allow unambiguous scientific communication.
|
ABAT,CASP7,IL1B,MEFV
|
SNP |
rSNP ID number, or “rs” ID comma separated. This is an identification tag assigned by NCBI to a
group
(or cluster) of SNPs that map to an identical location.
|
rs11139084,rs1217338,rs2781542,rs7866199
|
Imaging feature |
Imaging features included in NeuroImage
Feature
Terminology (NIFT)1.
|
CSF levels, white matter volume, frontal lobe volume, hyperintensity |
1. Iyappan A., Younesi E., Redolfi A., Vrooman H., Khanna S., Frisoni G., Hofmann-Apitius M. For the
Alzheimer'Disease NeuroImage Feature Terminology (NIFT): A controlled terminology for the annotation of
brain imaging features. 2017, accepted.
Enriched subgraphs and multimodal linked data
Below, the results page where multiple subgraphs can be selected, see the enrichment scores, import results
and have access to
all cross scaled data (co-expressed genes, drugs, miRNA or related clinical
trials). Information can be exported by clicking in
"Download Excel" button located at the bottom
of the page. In order to access the networks, simply click in
"Visualize Network". In the top-right
of the page, an
info icon contains the legend of result page.
Network visualization and functionalities
Once in the network visualization, exporting options (BEL, image, json) are available as well as search
funcionalities for edges and nodes. By clicking in "functionalities" you can visualize and find specific
paths and exported to BEL. It is also possible interact with the subgraph by expanding the knowledge around
nodes or deleting the nodes from the graph.
Candidate mechanism retrieval
Users should select candidate nodes based on their interest. From all data-mapped nodes to this selected
node, candidate mechanisms (represented as a chain of causation or paths) are displayed in the "Candidate
mechanism" tab.
Clicking in each candidate mechanism, allows you to navigate and visualize it in the interface. An example
of how candidate mechanism are displayed is shown below. This simple example has been generated by
submitting a geneset of interest in the IMI project AETIONOMY: FOXA2, TH, BCL2L1, NGF in the context of
Parkinson's. Later,
all suggested subgraphs were selected. In the visualization site, "alpha synuclein toxicity" was selected as
a biological process of interest in the disease. All paths from each of the nodes in the network
which map the data-mapped nodes to the interesting node "alpha synuclein toxicity" are displayed in
different colors (screenshot below). Here, PITX3 is suggested to be the key player in this particular
mechanism.