[Jalview-discuss] losing stop codons

[Juan Carlos Sanchez]

Hi,
I noticed that when having stop sign in an aminoacid sequence "*" , when using MUSCLE web service to realign the sequences it will be converted to a gap "-"
I know is not common to align with stops, but in some cases I think it happens.
It might be an issue of MUSCLE itself, but I thought it could be interesting to share this bug here.

Cheers
jc

[Jim Procter]

Hi JC - :)  Thanks for the interesting email !

On 03/07/2018 03:27, Juan Carlos Sanchez wrote:
> I noticed that when having stop sign in an aminoacid sequence “*” ,
> when using MUSCLE web service to realign the sequences it will be
> converted to a gap “-“
Ah - that sounds like Muscle is just treating '*' as a gap.
> I know is not common to align with stops, but in some cases I think it
> happens.It might be an issue of MUSCLE itself, but I thought it could
> be interesting to share this bug here.
What would be the ideal behaviour  ? 

We could, for instance, have Jalview reinsert the '*' in the alignment
that is returned from Muscle...

Jim.

--
-------------------------------------------------------------------
Dr JB Procter, Jalview Coordinator, The Barton Group
Division of Computational Biology, School of Life Sciences
University of Dundee, Dundee DD1 5EH, UK.
[phone] | www.jalview.org | www.compbio.dundee.ac.uk

[Juan Carlos Sanchez]

Hi Jim,
thanks for the quick reply!
i dont know if reinserting the '*' would be enough, because if MUSCLE treat it as a gap '-', and by any chance eliminate the position when making an alignment, just reintroducing it after would mess up the result.
Do you think these two options could be taken into account?
Also i havent had time to check if MUSCLE behaviour could be fix changing some default parameter...
cheers
jc

On 3 Jul 2018, at 7:08 pm, Jim Procter <***@compbio.dundee.ac.uk<mailto:***@compbio.dundee.ac.uk>> wrote:

Hi JC - :) Thanks for the interesting email !

On 03/07/2018 03:27, Juan Carlos Sanchez wrote:
I noticed that when having stop sign in an aminoacid sequence “*” , when using MUSCLE web service to realign the sequences it will be converted to a gap “-“
Ah - that sounds like Muscle is just treating '*' as a gap.
I know is not common to align with stops, but in some cases I think it happens. It might be an issue of MUSCLE itself, but I thought it could be interesting to share this bug here.
What would be the ideal behaviour ?

We could, for instance, have Jalview reinsert the '*' in the alignment that is returned from Muscle...

Jim.


--
-------------------------------------------------------------------
Dr JB Procter, Jalview Coordinator, The Barton Group
Division of Computational Biology, School of Life Sciences
University of Dundee, Dundee DD1 5EH, UK.
[phone] | www.jalview.org<http://www.jalview.org> | www.compbio.dundee.ac.uk<http://www.compbio.dundee.ac.uk>



_______________________________________________
Jalview-discuss mailing list
Jalview-***@jalview.org<mailto:Jalview-***@jalview.org>
http://www.compbio.dundee.ac.uk/mailman/listinfo/jalview-discuss