Hi,

I have a problem with the REL analysis on Datamonkey.  If I find recombination using GARD and then try to use the GARD trees for the REL analysis, the result ignores the breakpoints found by GARD and doesn't partition the data set.  This does not happen when I do PARRIS or FEL and try to use the GARD trees.  Any ideas what the problem is?

Andrew

Hi Andrew,

OOPS -- this bug must have been around for years and you are the first one to report it! Thanks.
Could you try again, the REL analysis should use GARD trees if requested.

Sergei

Hi Sergei,

I tried it again and it works, so thanks for the quick fix!  When the job is in the queue it still says there is only one partition, but the results look like it did do the further partitions.  

I found one other issue with the model selection.  After I do the model selection algorithm and start one of the other tests, Datamonkey seem to copy over the best model but only partially.  The top row of values is usually right but the second row is less often right and the final "GT" value is usually wrong.  I can always insert the correct model, so I can get the correct analysis, but is the correct model supposed to copied to the various tests' set-up screens?

Andrew

Hi Andrew,

Are you using IE? I need to debug some weirdness with Javascript and Internet Explorer.
(I fixed the JS in IE. Apparently whereas every other browser understands "string[index]" to mean the character at position index+1 in the string, IE forces one to use string.charAt (index)).

Sergei

PS Good catch with the queue status line. I fixed that.
PP

Hi Sergei,

I've been using Seamonkey, although I could use IE if needed.

Andrew

Hi Andrew,

Shouldn't matter now -- I think I fixed the IE-specific issue.

Sergei