[ssml] FW: Evolutionary pathway from seq alignment

[Dan Bolser]

On Tue, 24 Aug 2004, Goel, Manisha wrote:

>
>
>>  -----Original Message-----
>> From: 	Goel, Manisha  
>> Sent:	Tuesday, August 24, 2004 2:29 PM
>> To:	'pdb-l at sdsc.edu'
>> Subject:	Evolutionary pathway from seq alignment
>> 
>> Dear All,
>> 
>> I am quite new to sequence alignments and have a basic question
>> (nevertheless I can't figure it out for myself), so please help.
>> 
>> I am using a variety of seq alignment programs to align the output of
>> PSI-Blast. 
>> In a particular program my query is among the top sequences and the
>> rest of the hits with decreasing similarity arranged below.
>> In the other program, a few seq not so very similar to the query are
>> at top, and then a stack of seq similar to query (including query) are
>> arranged in the middle followed by other seqs arranged in order of
>> lowering similarity.
>> 
>> Does this mean that the two programs are showing two different
>> evolutionary pathways for the sequences picked up by PSI-Blast.

Are you talking about trees produced by the alignment programs? or just
the order of the sequences in the alignment? Generally their is no
absolute meaning to the order of the sequences in an alignment - different
programs will use specific or random orderings depending on more or less
arbitary decisions. 

If you make a phylogenetic tree from your alignment you get a better
picture of the 'evolutionary pathway' of your sequence. It will be linked
to closly related sequences on close branches and distantly related
sequences on distant branches. However, you should learn more about
phylogenetics before interpreting your trees (as it isn't always simple:)

So you have three stages to think about

Query sequence
     |
1) Find homologues
     |
2) Align homologues
     |
3) Build tree

1,2 and 3 all have different programs and their own problems. The best way
to learn is to try some, read the manuals and the literature :)

Best of luck!

>> 
>> Thanks for any enlightments !
>> - Manisha
>> 
>