> As long as what you show me are two reads and not the reference and a read, > I'm fine. I don't like it, but I've learned that whatever clever trick one > applies, there will always be a combination of mapping sequence and > sequencing > errors that will lead to situations like this. You are right, the situtation occurs at homopolymer frameshifts in the reference, so exactly as you describe below. Sorry for being slow :) > I have an idea how this kind of effect could be toned down a bit and the > pieces for it are almost in place in MIRA, but I need to test it (and hope it > really works). That sounds great! B On Fri, 11 Dec 2009 19:49:48 +0100 Bastien Chevreux <bach@xxxxxxxxxxxx> wrote: > On Freitag 11 Dezember 2009 Björn Nystedt wrote: > > I am still somewhat confused about exactly how the second error can occur. > > The situation is that this is what you would expect > > ...CATTTTTTC... > > ATTTTTTC... > > but this is what you get > > ...CATTTTTTC... > > A*TTTTTTC... > > I can see the point of relaxing gap penalties at ends of reads, but to > > prefer a mismatch over a match seems strange to me. Note also that there > > was no IUPAC in the reference in this case! > > As long as what you show me are two reads and not the reference and a read, > I'm fine. I don't like it, but I've learned that whatever clever trick one > applies, there will always be a combination of mapping sequence and > sequencing > errors that will lead to situations like this. > Welcome to the wonderful world of mapping. > > Imagine you have a backbone: > > B ...GCTTTTTTCG... > > You now map the first read which has one sequencing error right at the > beginning/end (an A instead of a C): > > B ...GCTTTTTTCG... > r1 ATTTTTTCG... > > Now you map a second read, where the sequecning error is an insertion (an > additional A) somewhere in the middle of the read. Remember, you map against > the reference: > > B ...GC*TTTTTTCG... > r1 A*TTTTTTCG... > r2 ...GCATTTTTTCG... > > Wham! You have the situation you described. From there on, things can only > deteriorate with additional reads :-) > > I have an idea how this kind of effect could be toned down a bit and the > pieces for it are almost in place in MIRA, but I need to test it (and hope it > really works). > > Regards, > Bastien > > -- > You have received this mail because you are subscribed to the mira_talk > mailing list. For information on how to subscribe or unsubscribe, please > visit http://www.chevreux.org/mira_mailinglists.html -- ==================================== Björn Nystedt, PhD Molecular Evolution EBC, Uppsala University Norbyv. 18C, 752 36 Uppsala Sweden phone: +46 (0)18-471 45 88 email: Bjorn.Nystedt@xxxxxxxxx ==================================== -- You have received this mail because you are subscribed to the mira_talk mailing list. For information on how to subscribe or unsubscribe, please visit http://www.chevreux.org/mira_mailinglists.html