v0.9

Summary : Using any clinical trial id, from across the globe, find the associated diseases/interventions, research articles and genes. Also discover relationships b/w various medical topics through co-occurrences in articles. Query the graph using SparQL from cli or GraphQL using any API client tool ex: Postman or curl

Feature list :

• Using GraphQL API knowledge graph can be queried using any API client tool ex: curl or Postman.
• Graph includes trials from across the globe. Data is sourced from WHO's ICTRP and clinicaltrials.gov
• Links from trial to MeSH vocabulary are added for conditions and interventions employed in the trial.
• Links from trial to PubMed articles are added. PubMed's experts curate this metadata information for each article.
• Added MRCOC to the graph for the selected articles linked to clinical trials.
• Added PheGenI links i.e. links from phenotype to genotype as links between MeSH DUI and GeneID.
• Added SparQL query execution feature. Adding CLI mode. Adding a count SparQL query for demo.
• 5 co-existing bi-partite graphs b/w trial--> condition, trial--> intervention, trial --> articles, article --> MeSH DUIs, gene id --> MeSH DUIs together comprise this knowledge graph.

Changes in this release : Enabled GraphQL server mode of execution.

Artifacts:

• For graph and trial ids list refer to v0.8 release as there are no changes to them.

v0.8

Summary : Using any clinical trial id, from across the globe, find the associated diseases/interventions, research articles and genes. Also discover relationships b/w various medical topics through co-occurrences in articles. Query the graph using SparQL from cli or GraphQL using any API client tool ex: Postman or curl

Feature list :

• Using GraphQL API knowledge graph can be queried using any API client tool ex: curl or Postman.
• Graph includes trials from across the globe. Data is sourced from WHO's ICTRP and clinicaltrials.gov
• Links from trial to MeSH vocabulary are added for conditions and interventions employed in the trial.
• Links from trial to PubMed articles are added. PubMed's experts curate this metadata information for each article.
• Added MRCOC to the graph for the selected articles linked to clinical trials.
• Added PheGenI links i.e. links from phenotype to genotype as links between MeSH DUI and GeneID.
• Added SparQL query execution feature. Adding CLI mode. Adding a count SparQL query for demo.
• 5 co-existing bi-partite graphs b/w trial--> condition, trial--> intervention, trial --> articles, article --> MeSH DUIs, gene id --> MeSH DUIs together comprise this knowledge graph.

Changes in this release : Enabled GraphQL API on the knowledge graph using HyperGraphQL

v0.7

Summary : One can move from any trial, across the globe, to medical vocabulary of diseases/interventions, to research articles, to genes. Also discover relationships b/w various medical topics through co-occurrences in articles. Along with ability to query the graph using SparQL

Feature list :

• Graph includes trials from across the globe. Data is sourced from WHO's ICTRP and clinicaltrials.gov
• Links from trial to MeSH vocabulary are added for conditions and interventions employed in the trial.
• Links from trial to PubMed articles are added. PubMed's experts curate this metadata information for each article.
• Added MRCOC to the graph for the selected articles linked to clinical trials.
• Added PheGenI links i.e. links from phenotype to genotype as links between MeSH DUI and GeneID.
• 5 co-existing bi-partite graphs b/w trial--> condition, trial--> intervention, trial --> articles, article --> MeSH DUIs, gene id --> MeSH DUIs together comprise this knowledge graph.

Changes in this release : Added SparQL query execution feature. Adding CLI mode. Adding a count SparQL query for demo.

v0.6

Summary - One can move from any trial, across the globe, to medical vocabulary of diseases/interventions, to research articles, to genes. Also discover relationships b/w various medical topics through co-occurrences in articles.

• Graph includes trials across the globe from nearly 20 clinical trial registries. Data is sourced from WHO's ICTRP and clinicaltrials.gov
• Links from trial to MeSH vocabulary are added for conditions and interventions employed in the trial.
• Links from trial to PubMed articles are added. PubMed's experts curate this metadata information for each article.
• Added MRCOC to the graph for the selected articles linked to clinical trials.
• 5 co-existing bi-partite graphs b/w trial--> condition, trial--> intervention, trial --> articles, article --> MeSH DUIs, gene id --> MeSH DUIs together comprise this knowledge graph.

Changes in this release : Added PheGenI links i.e. links from phenotype to genotype as links between MeSH DUI and GeneID.

v0.5.1

• Summary - One can move from any trial, across the globe, to medical vocabulary of diseases/interventions, to research articles. Also discover relationships b/w various medical topics through co-occurrences in articles.
• Graph includes trials across the globe from nearly 20 clinical trial registries. Data is sourced from WHO's ICTRP and clinicaltrials.gov
• Links from trial to MeSH vocabulary are added for conditions and interventions employed in the trial.
• Links from trial to PubMed articles are added. PubMed's experts curate this metadata information for each article.
• 4 co-existing bi-partite graphs b/w trial--> condition, trial--> intervention, trial --> articles, article --> MeSH DUIs together comprise this knowledge graph.

Changes in this release : Adding MRCOC to the graph for the selected articles linked to clinical trials.

v0.5

• Summary - One can move from any trial, across the globe, to medical vocabulary of diseases/interventions, to research articles. Also discover relationships b/w various medical topics through co-occurrences in articles.
• Graph includes trials across the globe from nearly 20 clinical trial registries. Data is sourced from WHO's ICTRP and clinicaltrials.gov
• Links from trial to MeSH vocabulary are added for conditions and interventions employed in the trial.
• Links from trial to PubMed articles are added. PubMed's experts curate this metadata information for each article.
• 4 co-existing bi-partite graphs b/w trial--> condition, trial--> intervention, trial --> articles, article --> MeSH DUIs together comprise this knowledge graph.

Changes in this release : Adding MRCOC to the graph for the selected articles linked to clinical trials.

Look at v0.5.1 for release artifacts and few fixes

v0.4

Features as on current release - 0.4

• Summary - One can move from any trial across the globe to medical vocabulary of diseases and interventions to research articles
• Graph includes trials across the globe from nearly 20 clinical trial registries. Data is sourced from WHO's ICTRP and clinicaltrials.gov
• Links from trial to MeSH vocabulary are added for conditions and interventions employed in the trial.
• Links from trial to PubMed articles are added. PubMed's experts curate this metadata information for each article.
• Three co-existing bi-partite graphs b/w trial--> condition, trial--> intervention, trial --> articles together comprise this knowledge graph.

Release Notes:

• List of trial ids to be incrementally bounced against Entrez API to generate the necessary incremental mappings b/w trials and PubMed articles

$ grep "Pubmed_Article" data/open_knowledge_graph_on_clinical_trials/vaidhyamegha_open_kg_clinical_trials.nt 
<https://clinicaltrials.gov/ct2/show/NCT00400075> <Pubmed_Article> "25153486" .
<https://clinicaltrials.gov/ct2/show/NCT03934957> <Pubmed_Article> "34064657" .

Next steps

• Symptoms - Diseases links using all PubMed Articles from MRCOC
• Symptoms - Genotype links using Phegeni

v0.3.1

• Adding links between trials and interventions in addition to trials and conditions.
• conditions and interventions are fetched from database (instead of files). Corresponding edges b/w trials and conditions, trials and interventions are added to RDF. For example :

  <https://clinicaltrials.gov/ct2/show/NCT00093782> <Condition> <http://id.nlm.nih.gov/mesh/2022/T000687> .
  <https://clinicaltrials.gov/ct2/show/NCT00093782> <Intervention> <http://id.nlm.nih.gov/mesh/2022/T538652> .

• All global trial's - 755450 - are added to RDF. For example :

<https://clinicaltrials.gov/ct2/show/NCT00172328> <TrialId> "NCT00172328" .
 <https://www.who.int/clinical-trials-registry-platform/CTRI/2021/05/033487> <TrialId> "CTRI/2021/05/033487" .

• Starting with a fresh model for final RDF instead of MeSH RDF. MeSH ids that are not linked to any trial are not considered. This reduces the graph size considerably.
• Trial records are fetched from ICTRP's weekly + periodic full export and AACT's daily + monthly full snapshot.
• Trials are written down to a file (will be used later) : vaidhyamegha_clinical_trials.csv.gz

  $ cat data/open_knowledge_graph_on_clinical_trials/vaidhyamegha_clinical_trials.csv | sort -u | wc -l
755450

• Download the RDF : vaidhyamegha_open_kg_clinical_trials_v0.3.nt.gz

v0.2

• all MeSH terms are retained as-is.
• Clinical trials from clincialtrials.gov are added.
• clinical trials are linked to the MeSH terms that correspond to the conditions on which the trial is/was conducted.