TOOLS

SemMedDB Database Download

We have deleted all previous versions of the SemMedDB database (versions 31, 40, 41, and 42).

The databases downloadable here (from version 3.0 on) have been created using the new database schema (The schema and database information is available here).

Note that the database file named "WHOLEDB" contains the entire database except the ENTITY table, which the new database provides. The individual tables are also provided separately. Users can download the entire database at once or the individual tables separately depending on their needs. The file names consist of four parts: database name _ R (or A) _ table name _ PubMed to date. The letter R represents that the database was generated with standard SemRep options, whereas A denotes that it was generated with the anaphora resolution option.

The new database schema differs from the previous one (versions 2X) in the following ways:

  1. We simplified the schema significantly by removing the CONCEPT, CONCEPT_SEMTYPE, PREDICATION_ARGUMENT, and SENTENCE_PREDICATION tables. The relevant contents of these tables can still be derived from PREDICATION if needed.
  2. A GENERIC_CONCEPT table has been added to the schema. This table contains generic concepts, as indicated by SemRep. The concepts that are not in this table are considered novel.

We no longer produce an annual release of the database of predications generated by SemRep using the sortal anaphora resolution. For sortal anaphora resolution in SemRep, see our BMC Bioinformatics paper.

The GENERIC_CONCEPT table has been updated in the June 30 2018 and all subsequent releases. Consequently, the SUBJECT_NOVELTY and OBJECT_NOVELTY columns of the PREDICATION table have been updated as follows: If the concept is not in the GENERIC_CONCEPT table, the value is set to 1; otherwise, it is set to 0.

Starting with version VER40, the PMID column in SENTENCE is dependent on the PMID in the CITATION table with a foreign key constraint. Therefore, all the PMIDs in the SENTENCE table have corresponding rows in the CITATION table, which has metadata PMID information.


Please note that all downloads in the tables below require a UMLS Terminology Services (UTS) account; to sign up for a UTS account, please click here.

Database name: semmedVER43_R (Processed using MEDLINE BASELINE 2021 + Covid-19 citations + PubMed Update Files through Aug 18, 2021) NEW

Semrep version: Regular semrep version 1.8
Number of citations processed: 32,818,567
Number of predications: 113,267,090
* This database was obtained from SemRep results with the anaphora feature turned off.

The next table contains links to MySQL download files.

TABLE NAME Size # Rows Download link sha1sum md5sum
CITATIONS 1545   32,963,356   download download download
ENTITY 38G   1,600,754,313   download download download
GENERIC_CONCEPT 4.7K  259  download download download
PREDICATION 2.7G   113,863,366   download download download
PREDICATION_AUX 3.5G   113,863,364   download download download
SENTENCE 14G   223,727,931   download download download
Entire Database 20G  N/A  download download download

 


The next table contains links to CSV files (there is no CSV file for the Entire Database).

TABLE NAME Size # Rows Download link sha1sum md5sum
CITATIONS 155M   32,963,356   download download download
ENTITY 40G   1,600,754,313   download download download
GENERIC_CONCEPT 3.9K  259  download download download
PREDICATION 2.8G   113,863,366   download download download
PREDICATION_AUX 3.7G   113,863,364   download download download
SENTENCE 14G   223,727,931   download download download