SemMedDB_UTH Database Outline

SemMedDB_UTH was created by the School of Biomedical Informatics at the University of Texas-Houston. The database holds semantic predications extracted from MEDLINE citations with a version of SemRep that increases the number of predications on drug therapies for cancer. The structure of SemMedDB_UTH is similar to the original SemMedDB provided by NLM.

Semantic predications in SemMedDB_UTH were extracted using SemRep (v1.5), with -D flag removed and keeping the -S flag. The number of predications extracted has been increased by modifying another NLM resource used by SemRep, MetaMap (v2013). MetaMap has been modified to use a special version of the UMLS Metathesaurus (v2013AB) in which NCI/CCN entries have been unsuppressed and marked as preferred terms. Only English language terms are used. In addition to these core changes, a new Entity table was added. It contains information for all entities extracted by MetaMap, not just those used in SemRep predications.

For details about the SemMedDB_UTH schema, click on the "Database Details" under SemMedDB_UTH menu.

To download the SemMedDB_UTH, a UTS account is required. Users that already have one need to click here and provide their UTS account information. To get a UTS account, click here.

For more information about SemMedDB_UTH, contact the following researchers.


Recent Publications
  1. Fathiamini S, Johnson AM, Zeng J, Araya A, Holla V, Bailey AM, Litzenburger, BC, Sanchez NS, Khotskaya Y, Xu H, Meric-Bernstam F, Bernstam EV*, Cohen T*. (2016). Automated Identification of Molecular Effects of Drugs (AIMED). J Am Med Inf Assoc, 2016 Jul;23(4):758-65. [* = co-senior/corresponding authors]