Skip to main content
Figure 1 | BMC Bioinformatics

Figure 1

From: Active learning for human protein-protein interaction prediction

Figure 1

Coverage and completeness of features. Each row corresponds to the feature vector for a protein pair. The feature vector has 27 elements corresponding to the various features such as Gene Ontology features, Gene Expression features, tissue feature etc. A feature element whose value is present is shown as a shaded box. A feature element whose value is missing is shown as a white box. Some vectors have very few missing values (high completeness). Typically these belong to interacting pair of proteins. On the other hand, many vectors have several missing values, and typically they correspond to negative (random) pairs. Some feature types (columns), such as tissue feature have low-coverage (many missing values), while some feature types such as gene expression have high-coverage (less missing values).

Back to article page