Hi Virginia ACT is using the percentage identities, score, match/alignment length reported by BLAST. So I guess you question is more a BLAST question but I will try and answer from an ACT point of view.
If you set the minimum % identity cut-off to zero then artemis will display all the results, i.e. nothing is filtered out. If you set the minimum size cut-off of 70 bp then only the matches with that number of base pairs or higher will be displayed. So you can use a combination of score, % identity and alignment size to define what you want to display and what may be meaningful for you to display. In very coarse terms the higher the % identity the more meaningful that hit may be. As the score comes from using a scoring matrix this cut-off takes into account the similarity as well and so you can also use this when filtering what is being displayed. Best Regards Tim Tim Carver The Wellcome Trust Sanger Institute Wellcome Trust Genome Campus, Hinxton, Cambridge, CB10 1SA, UK On 19/7/05 12:22 am, "Virginia Isabel Rich" <[EMAIL PROTECTED]> wrote: > hello there, > > many of us around our labs use Artemis and ACT with moderate frequency, > however > I'm up against an Artemis weirdness that no one around here can explain to me; > so I'm helping you folks can! > > In the Comparison View Pop-up Menu, it is possible to Set the Percent ID > cutoffs. This slides from 0% to 100%, and the default view is 0%. If I have my > match-size set at, say, 70 bp, what does it mean for the % identity to be 0?? > Logically, there's a 1-in-4 chance that a single base would match another > base, > so it seems like for anything 4bp or longer the true default identity cutoff > would be, say, 25%. Or if the cutoff is 50%, does that really mean that what > I'm seeing represented in Comparison View is 70-mers that share at least 35bp > with their matches? > > The more I think about this the more I think it's merely a BLAST-weirdness, > not > an Artemis one. What this % identity is really letting the user select is just > the % identity within whatever *aligned* sequences are present in the BLAST > match between the two genomes, and thus is not a particularly meaningful way > to > look, for example, at 80% IDENTICAL 70-MERS ACROSS TWO GENOMES (this really > being my goal). This is likely better achieved by determining what the e-value > for such a match across a 70-mer would be given these two genomes, and using > that to Set the Score Cutoff. > > Would you folks agree with this interpretation, and/or would you have any > suggestions for a better way to visualize genome comparisons via 70mers with > varying percents of identity? > > Thanks a lot! > > Virginia Rich > > ~~~~~~~~~~~~~~~~~~~ > Virginia Rich > [EMAIL PROTECTED] > DeLong Lab > MIT > > _______________________________________________ > Artemis-users mailing list > Artemis-users@sanger.ac.uk > http://lists.sanger.ac.uk/mailman/listinfo/artemis-users _______________________________________________ Artemis-users mailing list Artemis-users@sanger.ac.uk http://lists.sanger.ac.uk/mailman/listinfo/artemis-users
