redundancy filtering, memory, ...

Jalview Discussion
1

Hi!
I'm having a great time playing with Jalview Desktop, but I ran across a couple of things that may or may not be related.

Starting with a large alignment from Pfam (PF00532 full alignment with >5K sequences), which apparently includes many blank columns, I get different results when redundancy-filtering to 90% before and after removing the blank columns. Namely I get only a couple of sequences remaining when I filter before removing the blank columns, but hundreds remaining if I do it after removing the blank columns. Could it be due to running out of memory when I still have all the blank columns, or are blank columns somehow used when calculating % identity?

Suspecting memory issues, I got the launchApp jnlp for 1G from your website
http://www.jalview.org/jvmmemoryparams.html
However, when I use that and then from the Tools menu choose to show memory usage, the max value is no different than before. Shouldn't the max be 1G? Also, when I use launchApp it complains it couldn't open the examples file even though I had earlier specified in my Jalview preferences not to open that file. I'm a little unclear on whether it should be using the same settings (apart from memory allocation) as when I start by double-clicking the jalview.jnlp vs this launchApp file.

Thanks, as you know I'm a newbie but I try to figure it out from the documentation whenever possible!
Elaine

···

----------
Elaine C. Meng, Ph.D.
UCSF Computer Graphics Lab (Chimera team) and Babbitt Lab
Department of Pharmaceutical Chemistry
University of California, San Francisco

P.S. as to Mark's issue, I'm not having the problem with secondary structure display… I don't know what it looked like originally, but in what I downloaded today I see green arrows for strands and red oblongs for helices

2 
I'm having a great time playing with Jalview Desktop, but I ran across a couple of things that may or may not be related.

:slight_smile:

Starting with a large alignment from Pfam (PF00532 full alignment with >5K sequences), which apparently includes many blank columns, I get different results when redundancy-filtering to 90% before and after removing the blank columns.

yes. This is because Jalview isn’t excluding gapped columns from the redundancy calculation. Definitely an oversight (although it might sometime be useful). I also notice the documentation is pretty sparse on this function:

"Selecting the option in the Alignment window’s Edit menu or pressing ‘CONTROL+D’ brings up a dialog box asking you to select a threshold. If the percentage identity between the aligned positions of any two sequences in the visible alignment exceeds this value, the shorter sequence is discarded.

Note: The redundancy calculation is done when the dialog box is opened. For large numbers of sequences this can take a long time as all pairs have to be compared."

I’ll need to check exactly what’s going on here though, since I started to create some more rigorous tests for the PID function because I’ve been looking at implementing different measures. Hope I didn’t break it again !

Namely I get only a couple of sequences remaining when I filter before removing the blank columns, but hundreds remaining if I do it after removing the blank columns.  Could it be due to running out of memory when I still have all the blank columns, or are blank columns somehow used when calculating % identity?

nothing to do with memory. Just a slightly broken function that needs fixing!

Suspecting memory issues, I got the launchApp jnlp for 1G  from your website
 [http://www.jalview.org/jvmmemoryparams.html](http://www.jalview.org/jvmmemoryparams.html)
However, when I use that and then from the Tools menu choose to show memory usage, the max value is no different than before.  Shouldn't the max be 1G?  Also, when  I use launchApp it complains it couldn't open the examples file even though I had earlier specified in my Jalview preferences not to open that file.  I'm a little unclear on whether it should be using the same settings (apart from memory allocation) as when I start by double-clicking the jalview.jnlp vs this launchApp file.

this is a tiny but extremely annoying bug in the launchApp servlet that I’ve not gotten around to fixing.
Try this link instead: http://www.jalview.org/webstart/jalview_2G.jnlp

Thanks, as you know I'm a newbie but I try to figure it out from the documentation whenever possible!

Much appreciated !

Jim.

···

Hey there, Elaine !

On 08/07/2014 19:29, Elaine Meng wrote: