###Jill Mesirov - Broad Institute
@broadinstitute
Broad has generated >150 TBPs in a year, how was this number calculated?
###Bas Burger - BT Global Commerce
There were like 5 misspellings in the feedback from Siri in the demo..
(Pretty sure Siri doesn't tie in to 2rd party apps)
###David Dooling - Washington University
WUSTL
"When in doubt, throw it out"
Consensus is to store the BAM
[they] Track everything in great detail
"We try to elliminate people"
###Toby Bloom - Broad Institute
Another number: 3 TBases per day (≈ 4 TBytes)
QC feedback to the lab has become critical
Having data graphically displayed in real time has changed everything. (Uses Tableau: http://www.tableausoftware.com/)
80% at 20x
51 Illumina machines
(One particular day, 14 TBytes)
Network becomes bottleneck (storage you can scale)
Tried writing to named pipes in stead of lots and lots of files -> Improved turn-arund time by 50%.
Haven't looked much at Hadoop style stuff
Lots of work to make sure analysis is run on files in place, so users don't need to copy data.
###Chris Botka - Harvard Medical School
About 40TB of data with keywords "junk, trash, backup, copy etc"