Do any relations imply another relation?

[ejin700]

Do any of the relations in Hetionet imply the existence of another relation (that isn't the reverse relation)? For instance, there is a Disease - upregulates - Gene edge and a Disease - associates - Gene edge. In this case, does the Disease - upregulates - Gene edge imply that there should also be a Disease - associates - Gene edge?

From the documentation, I see that the data sources for the associates and upregulates edges are different, but from a biological perspective, if a disease upregulates a gene, does that mean it's also classified as being associated with that gene?

[dhimmel]

Do any of the relations in Hetionet imply the existence of another relation

No, although some are actually mutually exclusive like Compound–palliates–Disease and Compound–treats–Disease.

The best starting point for details on edges and their sources is the methods section here.

from a biological perspective, if a disease upregulates a gene, does that mean it's also classified as being associated

Disease–upregulates–Gene edges are pretty noisy so I'd be hesitant to dilute the signal in association edges with these more high-throughput and questionable edges. Users could decide to group edge types, but from a network construction perspective we chose to keep them separate.