About
Documentation
FAQ
Contact
|
Interaction Confidence Scoring Tool
|
|
|
1. Input interaction network
|
Select a file containing your network
(one interaction per row; file size limit: 3MB) |
|
| Does your file contain a header row? |
(select for yes) |
| Text delimiter |
|
|
|
| Alternatively, paste interactions here |
|
|
|
2. Methods and parameters
|
2.1. Topology-based methods
|
|
|
 Assigns weights to edges according to graphical co-clustering of their incident nodes.
Score range: [0, 1].
 Kamburov et al., BMC Bioinformatics, 2012
|
In case your network is weighted,
should interaction weights be considered? |
(select for yes)
|
| Optimal clustering inflation * |
estimate from random rewiring(s) with percent rewired links
use following inflation value:
|
|
|
Assigns weights to edges according to the number of common network neighbors of their incident nodes.
Score range: [0, 1].
Goldberg and Roth, PNAS, 2003
|
| Size of the genome under study (number of genes) * |
|
|
|
Embeds the network in a geometric space and weights edges according to the distance between their incident nodes in that space.
Score range: [0, 1].
Kuchaiev et al., PLoS Comput. Biol., 2009
|
| dim * (dimensionality of the target space) |
|
priorEdge * (prior belief about what fraction of all possible
pairs of proteins in the network really interact) |
|
priorNonEdge * (prior belief about what fraction of all possible
pairs of proteins in the network do not interact) |
|
d * (Gaussian mixtures to use while learning
P(dist|edge) and P(dist|non-edge) |
|
learnSetSize * (number of edges used for learning.
Should be less than the number of edges) |
|
| stopEps * (determines when to stop the EM algorithm) |
|
|
|
2.2. Annotation-based methods
|
|
|
Number of publications reporting each interaction. ConsensusPathDB is used as the literature annotation basis.
Score range: [0, ∞).
Kamburov et al., Nucleic Acids Res., 2011
|
| Interaction type * |
protein-protein interactions
genetic interactions
|
| Taxonomic species * |
|
| Accession number type * |
|
Also include PubMed IDs of the publications
reporting the interactions |
(select for yes)
|
|
|
Semantic similarity of the Gene Ontology annotation of interacting genes.
Score range: [0, 1].
Yu et al., Bioinformatics, 2010
|
| Taxonomic species * |
|
| Accession number type * |
|
| Sub-ontologies * |
biological process
cellular component
molecular function
|
| GOSemSim measure * |
|
| Drop codes |
|
Also include GO annotations
of interactors |
(select for yes)
|
|
|
Indicates whether interaction participants are found together in manually annotated pathways. ConsensusPathDB is used as the pathway annotation basis.
Score range: {0, 1} (discrete).
Kamburov et al., Nucleic Acids Res., 2011
|
| Taxonomic species * |
|
| Accession number type * |
|
Also include the names of pathways where
interaction participants are found together? |
(select for yes)
|
|
3. Integrate scores
|
Calculate aggregate scores
|
Integrates results from different scoring methods into an aggregate score through machine learning techniques. Score range: [0, 1].
Kamburov et al., Nucleic Acids Res., 2011
|
Select scores to integrate
(at least two) |
CAPPIC
Common neighbors
Geometric embedding
Literature evidence
GO semantic similarity (biological process)
GO semantic similarity (cellular component)
GO semantic similarity (molecular function)
Pathway co-occurrence
|
| Choose integration method |
|
Paste a subset of real interactions
from your input network |
|
Paste a subset of non-interacting gene/protein pairs
from your input network |
|
|
4. Submit
|
Type in the text from the image below
(click here if you cannot read the text in the image)
|
|
Type in your e-mail address if you would like
to be notified when results are ready |
|
|
|
|