MosaikAligner getting slower and slower?

kga1978 · 09-24-2011, 03:14 PM

Hi All,

I am using MosaikAligner to align reads to 50 different reference sequences one of the other (so the reads always stay the same). I am using the following parameters:

-hs 8 -act 14 -mmp 0.20 -m all -p 4

When I started this morning, each alignment would take ~ 6min, but as I have been eating my way through the different reference sequences (they are all the same length and very similar - 3500bp and ~20% divergent), MosaikAligner is getting slower and slower. Right now an alignment takes more than an hour (I have done 20 so far). I thought maybe it had to do with the computer I am doing it on, but I have restarted the computer, but to no avail. Also, I can see that the computer is using the four processors as predicted, leaving the other four processors unused (so nothing else is running in the background taking away power).

Any idea what is going on here?

Thanks,
Kristian

dsenalik · 09-25-2011, 09:41 AM

I have noticed that the time for analysis varies widely for different reference sequences. Some complete in seconds, some in minutes, and there are a few that take many hours. I have assumed it has something to do with coverage, or maybe repeat regions, but I never checked so this is just a wild guess.

So, after a while, you can be stuck with all slow ones, and the only solution is to just wait for them to finish. (Or buy a new computer with more CPU cores!)

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
how to speed up Mosaikaligner on Helicos reads?	feng	Bioinformatics	0	06-27-2011 07:39 PM
MosaikAligner: Low reads/sec	bre	Bioinformatics	4	03-19-2010 11:44 AM
MosaikAligner	hsmart	Bioinformatics	0	12-09-2009 05:51 AM
Memory problems with MosaikAligner 1.0.1307 (removed -km and -pm)	Magnus	Bioinformatics	0	11-04-2009 02:07 AM

09-24-2011, 03:14 PM	#1
kga1978 Senior Member Location: Boston, MA Join Date: Nov 2010 Posts: 100	MosaikAligner getting slower and slower? Hi All, I am using MosaikAligner to align reads to 50 different reference sequences one of the other (so the reads always stay the same). I am using the following parameters: -hs 8 -act 14 -mmp 0.20 -m all -p 4 When I started this morning, each alignment would take ~ 6min, but as I have been eating my way through the different reference sequences (they are all the same length and very similar - 3500bp and ~20% divergent), MosaikAligner is getting slower and slower. Right now an alignment takes more than an hour (I have done 20 so far). I thought maybe it had to do with the computer I am doing it on, but I have restarted the computer, but to no avail. Also, I can see that the computer is using the four processors as predicted, leaving the other four processors unused (so nothing else is running in the background taking away power). Any idea what is going on here? Thanks, Kristian

09-25-2011, 09:41 AM	#2
dsenalik Carrot Scientist Location: Madison WI USA Join Date: Nov 2009 Posts: 42	I have noticed that the time for analysis varies widely for different reference sequences. Some complete in seconds, some in minutes, and there are a few that take many hours. I have assumed it has something to do with coverage, or maybe repeat regions, but I never checked so this is just a wild guess. So, after a while, you can be stuck with all slow ones, and the only solution is to just wait for them to finish. (Or buy a new computer with more CPU cores!)