Network integration
Integration model
The package provided two algorithm for network integration, one is MashupIntegration
and another is GeneMANIAIntegration
(GeneMANIA's raw networks integration is not fully tested).
ModMashup.MashupIntegration
— Type.Modified Mashup algorithm for network integration. Inside MashupIntegration model, it contains all the result after mashup integration.
Fields
β::Vector
: Beta vector as a result of linear regression.
H::Matrix
: Rows of H represent patients embendding in networks.
net_weights::Dict{String, Float64}
: Normalized mean network weights
weights_mat::Matrix
: Columns of weights_mat is computed network weights for each round of cross validation.
cv_query::Matrix
: Columns of cv_query is query id for each round of cross validation.
singular_value_sqrt::Vector
: singular value from mashup for dimensianal reduction.
tally::Dict{String, Int}
: Network tally result
combined_network::Matrix
: Combined single similarity network using network weights.
Constructor
MashupIntegration()
Create empty MashupIntegration model.
ModMashup.GeneMANIAIntegration
— Type.GeneMANIA lienar regression algorithm for network integration.
Fields
net_weights::Dict{String, Float64}
: A dictionalry map network name to its final network weights result, which is same with GeneMANIA.jar.
combined_network::Matrix
: Combined single similarity network using network weights.
normalized::Bool
: Wether normlize the network weights.
reg::Bool
: Wether add regularization term to the model.
Generic integration method
We have one generic function network_integration!
to provide same interface for both mashup and genemania integration.
The database contains the input. After the computation, the result will be saved on the model.
ModMashup.network_integration!
— Method.network_integration!(model::MashupIntegration, database::GMANIA)
Implement modified mashup network integration. Result will be save in the model. See MashupIntegration
for more information about the result.
Arguments
model::MashupIntegration
: Mashup network integration model.
database::Database
: Store general information about the patients and networks.
ModMashup.network_integration!
— Method.network_integration!(model::GeneMANIAIntegration, database::GMANIA)
Arguments
model::GeneMANIAIntegration
: GeneMANIA network integration model.
database::Database
: Store general information about the patients and networks.
Implement Raw mashup network integration. Result will be saved in the model. See GeneMANIAIntegration
for more information about the result. (Currently not fully tested.)