Web display speed ups and flat file usage

We’re always looking at ways we can improve the Ensembl browser experience.  Quite often this results in a focus on the speed at which a particular display or track can be provided.  Over the past year or so we have been generating various pre-computed data which is compressed and optimised for web display purposes.  For release 62 we took the step of moving some of this data outside of the database into binary files.

The signal plot or ‘wiggle’ style displays provided as part of the regulation evidence (see ‘Functional genomics’ in the config panel) are now served from collection or ‘col’ files.  This provides significant speed ups, allowing many more tracks to be turned on without adversely affecting the response time of the display.  In fact it is now possible to turn on all the current signal plots for human, that’s 334 distinct data sets!

Moving the data outside of the database also overcomes various constraints and issues with managing the data and much reduces the size of the funcgen database.  This is beneficial for those who want to download the funcgen databases for API access only, for those running mirror sites if you want to display the signal plots you will need to down load the col files.

We intend to broaden support for standard data formats (e.g. BAM etc) as we identify more data which is amenable to flat file access.  As these data are not contained in the MySQL data dumps, we have created a new area on the FTP site:


More information can be found in the README files contained in this directory.