BWA memory issues

[tgenahmet]

We're trying to use BWA with paired end reads (bwa sampe with default settings) and running into memory problems when aligning to whole human genome. It maxes out our memory on a machine with 24GB of RAM even with 12K reads.

Anyone else having similar issues?

[lh3]

What is the length of your reads?

[tgenahmet]

We were running it with 36bp reads.

[lh3]

I have only tested bwa for SOLiD data on one run and it works fine. I am not sure why it costs so much memory for your data. What is peak memory usage? Could that be because your machine has other processes using a lot of memory?