View Single Post
Old 09-16-2010, 11:13 AM   #4
fkrueger
Senior Member
 
Location: Cambridge, UK

Join Date: Sep 2009
Posts: 622
Default

Hi cmawhinney,

according to our experience you are only likely to loose (up to a quite substantial amount of) clusters if you are using a low number of multi-plex tags. We also performed a cluster density - low complexity sequence simulation program which supports our empirical findings that if you have 4 or more different sequence tags at the start of your sequences you should not run into any problems. Using 10 different barcodes should definitely provide enough variability in the start, thus spiking in a PhiX sample is not necessary.

If you have any further questions about initial biased sequences feel free to send me an email, by now we have done quite a few rescues of runs which failed due to this reason.
fkrueger is offline   Reply With Quote