FunSimMat release notes

Changes in release 6.0

Source data: UniProtKB 2012_01, Pfam 26, SMART from InterPro release 35.0, and OMIM (February 22, 2011).

Source data: UniProtKB 2010_10, Pfam 24, SMART from InterPro release 28.0, and OMIM (October 12, 2010).

Source data: UniProtKB 2010_08, Pfam 24, SMART from InterPro release 27.0, and OMIM (July 15, 2010).

Source data: UniProKB 15.12, Pfam 24, SMART from InterPro release 24.0, and OMIM (January 13, 2010).

Source data: UniProKB 15.3, GOA release May 2009, Pfam 23, SMART from InterPro release 20, and OMIM (June 10, 2009).
"NaN" replaced by "NA": Instead of "NaN", "NA" is used to indicate missing similarity values.
FunSimMat now includes GO annotation from the GOA project at EBI.
This almost doubles the number of GO annotation.
Term Overlap (TO) and Normalized Term Overlap (NTO) measures: We added the recently published TO and NTO
measures to the database. By default, they are not shown in the results table, but can be made visible with
the drop-down boxes. For more details on these measures, please have a look at this page.
RestLike interface: It is now possible to pass all query parameters within the URL. For more details on these measures,
please have a look at this page
Visualization: Results can now be visualized using Mondrian or Medusa. More information can be found here.
MedSim: FunSimMat now provides an option for prioritization of disease gene candidates. The paper describing the method
is currently under review.
Speed improvements: Queries speeds are greatly improved. Comparisons against a taxon or the complete database should now only
take a few seconds.
General improvements to the user interface

Source data: UniProKB 13.3, Pfam 22, SMART from InterPro release 16.3, and OMIM (May 19, 2008).
New GO relationship types "regulates", "positively_regulates", and "negatively_regulates":
These relationship types are ignored for the calculation of semantic similarity values.

Source data: UniProKB 12.8, Pfam 22, SMART from InterPro release 16.2, and OMIM (February 05, 2008).
"NaN" values: Similarity values between entities without annotation and other entities result in "NaN" values instead of 0.0 in old releases. This helps to distinguish between cases
without annotation and cases without similarity. The funSim and rfunSim scores are calculated only if both entities have biological process and molecular function annotation. The funSimAll
and the rfunSimAll scores are only calculated if both entities are annotated with all three ontologies.
Annotation class ids: Annotation class ids are stable between different FunSimMat releases. Obsolete classes are removed from new releases and new classes receive a new unique id.