3 September 2013
- In v2.2, the protein ids have changed. We are working on mapping the old ids to the new ids. Until this is complete, please use the previous version, http://v2-1.mistdb.com if you need to work with the old ids.
- An alternative, more stable approach is to only store references to sequences using the gene locus.
26 August 2013
- New version 2.2 - badly needed genome update; majority of new genomes are incomplete. Many genomes available for various strains of the same species.
- Minor usability improvements (e.g. the list of genomes and the taxonomy browser is much more interactive and dynamic.)
- No more genome metadata (phenotype, gram-stain, disease, etc.); NCBI no longer provides this information in an easily accessible format.
30 December 2011
- Genome update
- Made a slight modification to the signal protein prediction algorithm. Originally, all domain predictions were considered regardless of any overlap with other domains. Now, when dealing with overlapping domain predictions, all overlaps that have a lower E-value by at least 3 orders of magnitude are ignored. This adjustment still provides for close matches, but helps rule out likely false positives by giving preference to significantly stronger domain predictions.
4 November 2011
- New version 2.1 - overhauled the database schema and reloaded the database from scratch. Contains all available genomes available through September.
- Because the database has been rebuilt from scratch, all the previous MiST identifiers are no longer valid with MiST2.1. To help with the transition, the old database and identifiers are accessible at http://v2-0.mistdb.com. We apologize for any inconvenience this may cause.
- Massively improved search speeds! The Sphinx search engine now services all domain and description searches resulting in reasonable search times.
- 12 November 2010
- Moved to dedicated servers which should improve performance. There are still some optimization issues that need to be dealt with. In particular, domain searches are very slow and often do not complete.