Well, I found that running SAET before any assembly pipeline (mapping or denovo) is extremely useful. It really improves base-calling, N50 and N95 are increased and the number of total contigs is decreased. However, even after SAET, it is still not adviceable to convert cs format into bs format for assembly. Furthermore, SAET is memory greedy and requires quiet some time to run. Maybe, this could help you reconsider supporting Solid data for hybrid assemblies using Mira. It is, of course, your call, Bastien. Best regards, Juan Montenegro 2011/9/9 Robert Bruccoleri <bruc@xxxxxxxxxxxxxxxxxxxxx> > ** > As far as I can tell, SAET looks for unusual subsequences in the reads > (similar to the concept of Proposed End Clips in Mira) and corrects them > back to the frequent subsequences that are very similar. > > > > Bastien Chevreux wrote: > > On Friday 09 September 2011 07:40:04 000.calabi.yau.000@xxxxxxxxxxxxxxxxxxx: > > > have you tried error correcting the data using: > > > > > > SOLiD™ Accuracy Enhancement Tool => > > > http://solidsoftwaretools.com/gf/project/saet/ > > > > > > It is often recommended on the ABySS assembler mailing list for SOLiD > data > > > prior to assembly. => > http://www.bcgsc.ca/platform/bioinfo/software/abyss > > > I haven't used SOLiD data before, but maybe this helps you. Maybe ABySS > is > > > also an alternative to Velvet in your case... > > > Now that's something potentially extremely useful and I can understand > why it is recommended. Actually that's the tool that might change my > position on supporting SOLiD ... has anyone here experience with the > usefulness of it? > > > B. > > > > >