Hi, Herai!
I think you should leave the minMatch at default of 2,
i.e. don't specify that at all. This is not what
you think. Instead probably put -minScore=25.
The blatSuite.34.zip you are using contains a 32-bit
ELF executable. You can try the file command on the blat:
file ./blat
[hgwdev:herai> file ./blat
./blat: ELF 32-bit LSB executable, Intel 80386, version 1 (SYSV), for
GNU/Linux 2.2.5, dynamically linked (uses shared libs), for GNU/Linux
2.2.5, stripped
So, 32-bit can only handle 4GB ram.
You will have to either blat separate
chromosomes in batches and combine and
filter the psls yourself,
or you'll have to get the source
for blat and compile it following
the instructions to make a 64-bit version
that is not limited to 4GB.
Using the -noHead option can be handy
when combining psl outputs from individual
chromosome runs. Then you can just cat
them all together, sort them, and filter
with pslReps.
Thanks for the detailed information!
-Galt
On Thu, 19 Mar 2009, [email protected] wrote:
? Dear Dr. Galt.
? I am grateful
to you for your help.
? Bellow, I will list the required
informations that you solicited in last e-mail:
- PC: 64 bits
architecture, AMD Atlon X2 dual core proc, 8 GB-RAM, Linux Fedora 9 64
bits;
- UCSC Blat:
http://genome-test.cse.ucsc.edu/~kent/exe/linux/blatSuite.34.zip
-
Genome database:?
ftp://ftp.ncbi.nih.gov/genomes/H_sapiens/Assembled_chromosomes/
????????????????
-only reference sequence genome, build 36.3;
? ? ?
? ? ? ? ?? -I downloaded fasta files of
1-22, X and Y chromossomos, and agrupated each one in a single file (fasta
format): approximatedely 3GB.
-UCSC Blat command line user:
???? #./blat hs_ref_chr_all.fas query.fas output.out
-t=dna -q=dna -stepSize=5 -minScore=0 -minIdentity=0 -minMatch=25
?
?? I got an error message related to: not enough
system memory. However, I certified that my system has 8GBRAM, and the
graphical interface was not loaded.
?? Thanks!!
???????? Herai.
??
By the way, normally
questions should be addressed to
[email protected]
Thanks for the specific info about your machine.
Can you also show me example commandlines that
you are
using to run BLAT?
Also, where did you get blat,
and where did you get the human chromosome data?
BLAT on DNA with -stepSize=5 for extra sensitivity
takes
about 6GB of RAM for the entire human genome.
But a
32-bit compiled blat won't be able to use more
than 4GB address
space.
Thanks!
-Galt
On Thu, 19 Mar 2009, [email protected]
wrote:
?
Dear Dr. Galt.
? I'me using USCS Blat standalone
version
and wants to load and use the interely human genome
in my experiments.
At
the moment, I doing the
experiments with isolated human chromosomes,
because my
computer can't load all the chromosomes in memory. Probably,
I
done some mistakes in the Blat configuration,
because? my computer is
good. It is a 64 bits
architecture, AMD Atlon X2 dual core processor,
with
8 GB-RAM, Linux Fedora 64 bits. Like the input data to Blat
program, I
using the human genome, in fasta format, that has
approximatedely 3GB.
How
can I? try to
run, perfectly, the Blat?
?? Thank you
very much.
?? Regards,
?????
Herai.
_______________________________________________
Genome maillist - [email protected]
http://www.soe.ucsc.edu/mailman/listinfo/genome
