Wellness KP

Description: The Multiomics Wellness KP, created and maintained by the Multiomics Provider, provides observational knowledge derived through computation of correlations between many blood analytes. The original multiomic dataset was created through a wellness study that monitored the blood analytes and general wellness of a cohort of individuals that were largely healthy. Edges in this KP represent significant correlations between pairs of analytes.

Example: The Multiomics Wellness KP shows that the level of mercury observed in blood is strongly (positively) correlated with the level of docosahexanoic acid (DHA), a lipid found in fish and fish oil supplements. This correlation is consistent with mercury intake from contaminated fish. The strength of the correlation (Spearman rho) is 0.527, with p-value of 3E-221, as observed among N = 3197 individuals. The correlation is stronger among males (rho = 0.557, N = 1310) than among women (rho = 0.505, N = 1885). There is also an age dependence, with stronger correlation among individuals younger than 35 (rho = 0.588, N = 540) than among those between 35 and 55 (rho = 0.532, N = 1723) and those older than 55 (rho = 0.479, N = 932).

Key methodologic metrics provided on edges: The KP presents, for each correlation, statistical metrics like the strength of the correlation (Spearman rho), the p-value (Bonferroni-corrected for multiple testing), the number of individuals that contributed to the calculation (N), and details on what stratification was used, if any.

Additional Links: https://github.com/NCATSTranslator/Translator-All/wiki/Wellness-KP

The graph currently uses biolink:correlated_with as predicate. Node categories include biolink:ClinicalFinding, biolink:Protein, and biolink:SmallMolecule.

• Modes of Access - Reasoner API v.#.#.#, SmartAPI registry, KGX - formatted Knowledge Graph
  • BioThings Explorer: https://biothings.ncats.io/multiomics_wellness_kp
  • SmartAPI: https://smart-api.info/ui/02af7d098ab304e80d6f4806c3527027

Use Cases- A sample TRAPI formatted query: {
"message": {
"query_graph": {
"edges": {
"e0": {
"predicate": "biolink:correlated_with",
"subject": "n00",
"object": "n01"
}
},
"nodes": {
"n00": {
"category": "biolink:ClinicalFinding",
"id": "LOINC:1759-0"
},
"n01": {
"category": "biolink:ChemicalSubstance"
}
}
}
}
}

Knowledge Sources Accessed -

• ISB's Wellness dataset