|
From: David E. (gringer) <dav...@mp...> - 2011-07-15 08:43:58
|
On 15/07/11 10:12, David Eccles (gringer) wrote: >> Works fine on my branch master. > And also on my local copy of your master on my MPI desktop, even when > running with just 1 processor, carrying out the same steps that failed a > day ago on my laptop: > > $ readSimulator/VirtualNextGenSequencer > ~/data-for-system-tests/phix/phix.fasta 0 200 10 500000 50 > phix_500k_1.fasta phix_500k_2.fasta > > $ ./code/Ray -p phix_500k_1.fasta phix_500k_2.fasta > > [i.e. only one processor] > > Number of contigs: 1 > Total length of contigs: 5384 > Number of contigs >= 500 nt: 1 > Total length of contigs >= 500 nt: 5384 > Number of scaffolds: 1 > Total length of scaffolds: 5384 > Number of scaffolds >= 500 nt: 1 > Total length of scaffolds >= 500: 5384 > >> Your distribution has 2 peaks. > > Yes, I did notice that. Not quite sure how that happened given that I'm > working off the same code. I'll fetch the input data that failed once I > get back home, and also see what happens with real phiX data off our > sequencer. ... and I just re-ran with a 5k run, got one successful run, and then tried again and got an unsuccessful run: $ readSimulator/VirtualNextGenSequencer ~/data-for-system-tests/phix/phix.fasta 0 200 10 5000 50 phix_5k_1.fasta phix_5k_2.fasta $ mpirun -np ./code/Ray -p phix_5k_1.fasta phix_5k_2.fasta [so same as your steps, except a smaller genome size] These inputs seem to have a single peak: http://user.interface.org.nz/~gringer/hacking/phix_5k_1.fasta http://user.interface.org.nz/~gringer/hacking/phix_5k_2.fasta http://user.interface.org.nz/~gringer/hacking/coverage_5k_bad.pdf -- David |
