Re: Bioinformatics: Slow Parsing of a Fasta File

Bioperl has considerable overhead, but your numbers are extreme. Maybe you run an old bioperl? ( perl -MBio::SeqIO -e 'print $Bio::SeqIO::VERSION, "\n";' )

I copied your sequences to a 200+ MB file, and tested with and without bioperl, and with a desktop vs server. The slowest I got was still within 5 minutes. The faster machine ran your program in just over 1 minute, the straight perl version in 2s.

(reader.pl is basically your code above).


##############
#
#  5-year old desktop machine
#

$ ls -lh Test.fasta 
-rw-r--r-- 1 aardvark aardvark 234M Jul 27 23:02 Test.fasta

$ time perl -ne "print if /^>/" Test.fasta | wc -l 
1572864

real    0m5.709s
user    0m4.279s
sys     0m1.233s

$  time perl ./reader.pl | wc -l
1572864

real    4m26.494s
user    4m8.936s
sys     0m1.607s


##############
#
#  server 
#

$ ls -lh Test.fasta 
-rw-rw-r-- 1 aardvark aardvark 234M Jul 27 22:51 Test.fasta

$ time perl -ne "print if /^>/" Test.fasta | wc -l  
1572864

real    0m1.774s
user    0m1.607s
sys     0m0.214s

$ time perl ./reader.pl | wc -l
1572864

real    1m9.258s
user    1m9.012s
sys     0m0.229s
[download]

Comment on Re: Bioinformatics: Slow Parsing of a Fasta File Download Code

Replies are listed 'Best First'.
Re^2: Bioinformatics: Slow Parsing of a Fasta File by Anonymous Monk on Jul 28, 2010 at 06:34 UTC
The version that I have is 1.006001, how old is that ? and if its rusting old what is the best way to make an update? `I am testing that right on the laptop itself so there's no issue of a +network hindrance of any sort.` [download]	[reply] [d/l]