Machine-learning approaches, such as neural networks, support vector machines, or Bayesian networks, typically combine protein-interaction data with other sources of information, such as protein sequence and structure, pathway membership, gene expression, or the genome-wide association study (GWAS) p values
