Skip to main content
Fig. 2 | BMC Bioinformatics

Fig. 2

From: GraphTar: applying word2vec and graph neural networks to miRNA target prediction

Fig. 2

In the process of preparing the input graph, we initially segment the sequences of both the mRNA and miRNA’s Minimal Binding Site (MBS) into triplets. Subsequently, the sequences are aligned based on their starting triplets. Each triplet constitutes an individual graph node, and the nearest nodes are linked by bidirectional edges to construct an undirected graph. This graph functions as the input for GraphTar. It’s important to mention that prior to the graph preparation procedure, the sequence’s triplets were transformed into dense vectors using CBOW models

Back to article page