[mira_talk] Re: Mira mailing list
- From: Sandrine Moreira <sandrine.moreira@xxxxxxxxxxxx>
- To: mira_talk@xxxxxxxxxxxxx
- Date: Mon, 10 Jan 2011 18:44:48 -0500
Le 2011-01-10 à 1:42 PM, Bastien Chevreux a écrit :
> On Montag 10 Januar 2011 Sandrine Moreira wrote:
>> The technical suport suggest they could either :
>> - Work on the parallelization of the code
>> - Implement the "checkpoint" functionnality (to stop and restart the
>> assembly after 7 days)
>
> Is that a help offer?
Well, if our development could help others ....
> Whomever wants to add to MIRA is very welcome, though
> I'll note that there's a reason these two functions are not fully implemented
> yet: they're not easy to implement and it'll at least a couple of weeks to
> implement (and test!) either one.
I'm sure that it's a tricky part, but we can still evaluate the feasability and
the difficulty.
We will certainly implement the checkpoint due to this time limitation on the
cluster (jobs longer than 7 days are killed).
For the parallelization, it's less probable but it was worth asking if somebody
has already developed half of the code ;-)
>
>> Before investing too much time on one of these solutions, does anyone begin
>> working on one (both ?!) of these points ? Bastien ?
>> Bastien, I read on the doc that you are planning to work on these points,
>> did you already do something in this direction ?
>
> Both points are currently on hold, I have a couple of other things with
> higher
> priority on my list atm. The basics for checkpointing are there (MAF is a
> result of it), but then I needed to shift focus when I saw that there was
> more
> to it than I wanted.
Ok, we will surely interact with you for some precise questions on the
checkpoint, if you don't mind.
Is there some code related to this part which is not commited on sourceforge ?
>
>> BTW, I would be also interested by a small dataset to test our
>> developements. If you have one to suggest ...
>
> For that I suppose that any bacterial Sanger data set from the NCBI trace
> archive should work pretty well. I'd have to search for one though, all the
> ones I had I lost a couple of weeks ago.
I would prefer 454 data to stay closer as possible as our own data.
BTW, I perhaps found a mitochondrial genome for my test : small but complex
enough, with high %AT and nice homopolymers errors.
Still need to have a closer look to see if it's really convenient.
SaM
>
> B.
>
> --
> 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
--
Sandrine Moreira Rousseau
Doctorante en Bio-informatique
(514) 343-6111 poste 2842
Centre Robert-Cedergren en Bio-informatique et Génomique
Université de Montréal
Montréal, Québec, Canada
--
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
Other related posts:
