[mira_talk] MIRA and gap penalty
- From: Laurent Abi-Rached <Laurent.Abi-Rached@xxxxxxxxxxx>
- To: mira_talk@xxxxxxxxxxxxx
- Date: Wed, 31 Oct 2012 10:31:20 +0100
Hello,
I have a question regarding how gap penalties are set in MIRA: indeed,
while conducting a batch of tests, I noticed that MIRA would sometimes
favor gaps over mismatches; for example, instead of generating the
alignment #1 below with two mismatches and no gaps, MIRA would create
alignment #2 with no mismatches but two gaps.
Alignment #1 (expected)
AGCCCCACGTGC CA TGGAGGGACATGGAA
AGCCCCACGTGC AG TGGAGGGACATGGAA
Alignment #2 (generated by MIRA)
AGCCCCACGTG CCA* TGGAGGGACATGGAA
AGCCCCACGTG *CAG TGGAGGGACATGGAA
Since these sequences are coding sequences, alignment #1 should be
favored but I have not been able so far to find settings that would lead
to such an alignment. In particular, I made sure egp was set to yes and
tried different values for egpl (2 and 10) but alignment #1 is still
favored.
I am thus wondering if there is a way to increase the gap penalty to
prevent MIRA from favoring alignment #2 over alignment #1?
In case this is useful, to conduct these tests I used the latest version
of MIRA (dev3.9.5) and the following parameters:
--job=mapping,genome,accurate,Solexa -CO:mr=0 -GE:not=6 -AS:nop=3:ard=0
-CL:pechsgp=1:pecbph=12:cbse=1 -SK:not=6:bph=12:pr=70 -MI:somrnl=0
SOLEXA_SETTINGS -CL:pec=0 -AL:mo=25:ms=25:mrs=70:shme=10:egp=1:egpl=10
-AS:mrl=20 -CO:msr=0:amgb=1:amgbemc=1:rodirs=10
Also, this mapping assembly with Solexa reads has a coverage of about
120X in the region where the problem above occurs (on a side note, a
small number of reads (7-8) are aligned 'properly' (with two mismatches)
but all the other reads have the first or the second gap).
Thanks,
Laurent
--
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:
