testing bwa with 454 hsap data: infinite loop on sam generation?

[drio]

I am running bwa (0.5.0 (r1200)) on a linux box:

2.6.18-128.4.1.el5 (64)
gcc version 4.1.2
glib 2.5

I run it like this:

ref="./indexes/hsap.fasta"
reads="./fastqs/small.fastq"
align_out="align.sai"

bwa index -a bwtsw $ref
bwa dbwtsw -t 8 $ref $reads > $align_out
bwa samse $ref $align_out $reads > $sam_out

The first problem comes after generating the reference. Bwa is trying
to find:

hsap.fasta.nt.amb
hsap.fasta.nt.ann
hsap.fasta.nt.pac

But the index command generated:

hsap.fasta.amb
hsap.fasta.ann
hsap.fasta.pac

THat is easy to fix, but I would like to know why is doing that.

The other issue is more problematic. When trying to generate
the sam, bwa seems to stay in a loop and it doesn't generate anything.
When monitoring the process, it seems it is reading from a filedescriptor:

0.000000 [ 3bef0c5610] read(5, "", 4096) = 0
0.000106 [ 3bef0c5610] read(5, "", 2467500032) = 0
1.751992 [ 3bef0c5610] read(5, "", 4096) = 0
0.000095 [ 3bef0c5610] read(5, "", 2467500032) = 0
1.753771 [ 3bef0c5610] read(5, "", 4096) = 0
0.000094 [ 3bef0c5610] read(5, "", 2467500032) = 0
1.756972 [ 3bef0c5610] read(5, "", 4096) = 0
0.000092 [ 3bef0c5610] read(5, "", 2467500032) = 0

What do you think is happening?

[lh3]

Firstly you should use bwa-0.5.4 and secondly, dbwtsw output sam. You do not need to run samse.

[drio]

The reason why I used 0.5.0 is because I went to the sourceforge download page via the maq site. If coming through there that's the latest version you get from source
forge.

Can you change your link in the maq site to point to http://bio-bwa.sourceforge.net/ ?

Thanks for the reply.