Table 1.
Qualitative comparison of ligand–receptor and integrative databases.
| Resource | Inter‐actions | Directed inter‐actions | Signs (positive/negative) | Transcriptional regulation | Intracellular pathways | Intercellular communication roles | Protein complexes | Integrative resource | Literature curated |
|---|---|---|---|---|---|---|---|---|---|
| Baccin2019 (e) | yes | yes (a) | no | no | no | yes (f) | yes | yes | yes (g) |
| CellCellInteractions | yes | yes (a) | no | no | no | yes (l) | no | yes | no |
| CellPhoneDB | yes | yes (a) | no | no | no | yes (d) | yes | yes | yes |
| ConsensusPathDB | yes | no | no | yes | yes | no | no | yes | yes (g) |
| EMBRACE (e) | yes | yes (a) | no | no | no | yes | no | yes (k) | yes (g) |
| HPMR | yes | yes (a) | no | no | no | yes | no | no | yes |
| ICELLNET | yes | yes (a) | no | no | no | yes | yes | no | yes |
| iTALK (h) | yes | yes (a) | no | no | no | yes | no | yes | yes (g) |
| Kirouac2010 | yes | yes (a) | no | no | no | yes | no | no | yes |
| LRdb | yes | yes (a) | no | no | no | yes | no | yes | yes (g) |
| PathwayCommons | yes | yes (m) | no | yes | yes | no | yes | yes | yes (g) |
| Ramilowski2015 | yes | yes (a) | no | no | no | yes | no | yes | yes (g) |
| SignaLink | yes | yes | yes | yes (i) | yes | yes | no | yes (j) | yes (g) |
| OmniPath | yes | yes (b) | yes | yes | yes | yes (c) | yes | yes | yes (g) |
OmniPath combines resources to build a network with directions and effect signs, including intra‐ and intercellular signaling, transcriptional regulation, and annotates proteins as ligands or receptors. Here, we show which of these features are covered by other databases: those specialized in ligand–receptor interactions and two large integrative network databases (ConsensusPathDB and Pathway Commons). (a) Implicit: if we assume always the ligand affects the receptor; (b) As in some of the constituent resources the directions are implicit, certain directions in the combined network are implicit; (c) Provides not only ligand and receptor annotation but further categories, for example adhesion, transporter, ECM, etc; (d) Apart from secreted (mostly ligand) and receptor provides a few further categories: integrin, collagen, transmembrane, peripheral, etc; (e) Data are for mouse, homology translation is necessary to derive human data; (f) For ligands, provides certain classification, e.g., cytokine, ECM, secreted, etc; (g) Only in part is literature curated; (h) Ligand–receptor interactions are classified as growth factor, cytokine, checkpoint, or other; (i) Contains transcriptional regulation but that part is not integrated by OmniPath; (j) OmniPath only integrates its original literature curation, not the secondary resources; (k) Only builds on Ramilowski et al; (l) Besides ligand and receptor only ECM; (m) Directionality information might be extracted from BioPAX.