TICTAC: Target illumination clinical trials analytics with cheminformatics

TESTING -- NOT YET VALIDATED.

Sex As Biological Variable: genes with expression profiles most differentiated by sex

TIN-X test

Counts of associations via UMLS for these sources: HPO, ICD9CM, ICD10CM, MedDRA, MeSH, OMIM, SNOMEDCT

Values are log10(n_nodes) in functional network for each gene-tissue pair, with minimum edge weight 0.5. n_genes = 1372, n_tissues = 145.

JensenLab vs Pubtator scores, PubMed NER, only those diseases where DO to MeSH mapping available and both scores are nonzero. DO hierarchy level is depth from root "disease", DOID:4.

DO and MeSH levels. DO level is depth from root "disease", DOID:4. MeSH level is number of fields in "Tree Number". Note that some treenums not available, hence level 0. This is due to MeSH non-disease supplementary IDs mapped to diseases, not included here.

JensenLab vs Pubtator scores, PubMed NER, only those diseases where DO to MeSH mapping available and both scores are nonzero. MeSH hierarchy level is number of fields.

JensenLab vs Pubtator scores, PubMed NER. CAVEAT: This pie chart is FUNDAMENTALLY FLAWED due to the hierarchical nature of DO and MeSH disease terms, since the meaning of counts and the information content of each NER depends on the level.

Pubtator vs JensenLab, gene-PubMed NER analysis. Both datasets downloaded August 2016, mapped to TCRD v3.0.7 (September 2016).

Pubtator vs JensenLab, gene-PubMed NER analysis. Both datasets downloaded August 2016, mapped to TCRD v3.0.7 (September 2016).

Pubtator vs JensenLab, gene-PubMed NER analysis. Both datasets downloaded August 2016, mapped to TCRD v3.0.7 (September 2016).

Pubtator vs JensenLab, gene-PubMed NER analysis. Both datasets downloaded August 2016, mapped to TCRD v3.0.7 (September 2016).