Advertisement
Table 3 Common knowledge representations desired the transition
| Analytical Function | Knowledge representations/interactions needed for the transition to causal modeling | Available? | Easy to view/anticipate using? |
|---|---|---|---|
| Contextualizing | Expression values from one's own experimental data as cues about regulatory processes and paths | ✓ | Yes, but not found |
| Indirect relationships and paths between proteins in networks of interactions | ✓ | Required time-consuming filtering | |
| Functional pathways related to sets of interactions | ✓ | SAGA, but not inter-active; or side by side | |
| Homology or pathway comparisons between species to see if molecules or interactions are conserved | ✓ | Yes | |
| Ability to filter by multiple variables at once or to filter to only the shared interactors between specified genes | No | No; strings in a field | |
| Test statistics | No | No | |
| Detail | Types of interactions: Physical binding, activation/inhibition, family member interactions, transcription/expression, translocation/secretion, phosphorylation | Partly | Buried in NLP free text; incomplete knowledge |
| Types of molecules: Distinctions between genes, proteins, chemical effectors DNA, mRNA, protein complexes, mRNA, enzyme | Partly | Incomplete knowledge | |
| Experiment type | ✓ | Strings in a field | |
| Ability to color code by attributes | ✓ | Time consuming |
