Skip to main content
Fig. 5 | BioData Mining

Fig. 5

From: A maximum flow-based network approach for identification of stable noncoding biomarkers associated with the multigenic neurological condition, autism

Fig. 5

Simplified depiction of graph model. Consider a simplified representation of the dataset, consisting of variants 1A, 1B, 1C, and 1D (located on chromosome 1) as well as variants 2A, 2B, 2C, and 2D (located on chromosome 2). Assume that a feature matrix consisting of all eight variants served as input to an L1-regularized classifier, which performed five-fold cross validation and identified a set of top-ranked variants for each validation fold. In the depiction above, we see that the variants are clustered by fold, with an edge connecting a pair of nodes n1 and n2 if they are located in neighboring folds and have an R2 value greater than 0.8. Note that identical variants (such as variant 1B in folds 4 and 5) are trivially in LD

Back to article page

BioData Mining

ISSN: 1756-0381