Dear all,
As far as I understand it, there are all sequences in a dataset compared to each other during construction of the multiple alignment by ClustalW. Is there a possibility to get a list of %-identities generated during the pairwise alignments? Thanks!
Petr
As far as I understand it, there are all sequences in a dataset compared to each other during construction of the multiple alignment by ClustalW.
That’s roughly correct - though it depends on the parameters passed to clustalW as to how the comparison is done.
Is there a possibility to get a list of %-identities generated during the pairwise alignments? Thanks!
Jalview doesn’t recover any of the pairwise scores generated by the ClustalW service at the moment - the JABAWS service doesn’t provide them, and I’m slightly embarrassed to say that there is also no menu option in Jalview 2.7 that allows you access to the matrix of pairwise percentage identities computed when you ask it to build a percentage identity based tree. You do get a pairwise score if you pick two or more sequences and select ‘Pairwise alignments’ - but the alignment report isn’t particularly easy to read if you have lots of sequences.
If you need it desperately, its fairly easy to create a groovy script that would dump pairwise percentage identities, or you could head over to the ClustalW server at EBI, run the alignment again, and download the similarity matrix that is produced along with the alignment and guide tree.
Sorry about that - it always seems I have something I need to do before I can create the ‘export pairwise similarity matrix’ menu item 
Jim.
On 30/05/2012 14:54, Petr Efler wrote: