I am using bowtie to align Illumina GAII sequences to two different refrerences. One is contining large contigs and have about 20 000 entries (> in fasta file about 1.4 Gb) and the other contains over 3 million entries (> in fasta file, about 2.3 Gb). Now the problem is that when I try to align reads to the large index, it takes a long time, by that I mean several hours, much longer than it takes to align to the human genome which is bigger in terms of Gb.
I also check the CPU usage, when I run against the human genome or my smaller index, user cpu is at 100% but with the big index, system CPU is running at 12% or similar. I checked with my sys admin and he said that bowtie was waiting for a system call "read" and it looked like it was reading one byte at a time.
changing the offrate to e.g. 3 from 5 makes it worse but memory is NOT exhausted in either case (total memory is 72GB). Does anyone have similar problems? Any suggestions?
I also check the CPU usage, when I run against the human genome or my smaller index, user cpu is at 100% but with the big index, system CPU is running at 12% or similar. I checked with my sys admin and he said that bowtie was waiting for a system call "read" and it looked like it was reading one byte at a time.
changing the offrate to e.g. 3 from 5 makes it worse but memory is NOT exhausted in either case (total memory is 72GB). Does anyone have similar problems? Any suggestions?