Text mining with Linked Life Data notes

Slide 1

(no notes)

Slide 2

LLD is a massive RDF warehouse that semantically integrates more than 25 biomedical data sources covering the full information path from: genes, proteins, molecular interactions, targets, drugs, disease patient. (green boxes of the upper-right diagram)

This knowledge base is combined with thesauri and textual information (blue boxes) from UMLS, OBO foundry, Pubmed and ClinicalTrials.gov

Slide 3

LLD is excellent tool for creating semantic annotations. Semantic annotations are textual annotation that points to an ontology identifier (URI).

Slide 4

SKOS is a data model that allows the simple organisation of the thesauri. It gives you a standard way to express broader/narrower concept, their synonyms, but also to publish it on the web.

(Describe the diagram – the graph structure on the right represent an example that describes Asthma, all specialization/generalizations of the disease)

Slide 5

SKOS gives excellent capabilities for aligning multiple concept hierarchies

Slide 6

This is the view of the Asthma concept in a web browsed. The page is offered by LLD service.

Slide 7

You can see a more practical way to see SKOS schema. All data is coming from more than 120 thesauri integrated using UMLS data source. Definition window = a formal definition of the concept Concept labels = alignment between all synonyms across the different thesauri Concept types = classification of the concept in the UMLS semantic network You can download the data in JSON, RDF, N3, NTriples

Slide 8

The data in LLD and described by the SKOS schema is an excellent tool for getting specific sets of concepts and populating gazetteer lists.

Slide 9

Here is an example query that you can use with LKB gazetteer to get all respiration disorders. Just choose any concepts from the LLD repository and replace it’s prefLabel (title) with “Respiration Disorders”

Slide 10

Simple guide how to do it. It is important to point back that all annotations have class and inst annotations

Slide 11

(no notes)