Friday, October 29, 2010

running pace new method

Step 1. generate datafile with n sequences

head -n*2 original.fsa > small_sample.fsa

Step 2. run preprocess

./preprocessPaCE 5k.fsa 1

Step 3. run GItitle.pl

./GItitle.pl 5k.fsa.PaCE > 5k.GI.fsa

Step 4. run PaCE

#!/bin/csh
#@ job_type = parallel
#@ class = LONG
#@ account_no = NONE
#@ node = 2
#@ tasks_per_node = 4
#@ checkpoint = no
#@ wall_clock_limit = 05:00:00
#@ error = $(Executable).$(Cluster).err
#@ output = $(Executable).$(Cluster).out
#@ environment = COPY_ALL
#@ queue

llmachinelist
mpirun -v -np 8 -machinefile /tmp/machinelist.$LOADL_STEP_ID ./PaCE_v9
/N/gpfs/cap3/leesangm/HumanMRNA/split1mil/5k.GI.PaCE 5000
./Phase.cfg



llsubmit submitTest.sh


Step 5. split PaCE output

./PaCEclusterFasta.pl test.fsa estClust.5000.3.PaCE


#####
Notes


1. Pace: How to run

mpirun -v -np 8 ./PaCE /N/gpfs/cap3/leesangm/HumanMRNA/example5K/5k.GI.fsa 5000 ./PaCE.cfg

No comments:

Post a Comment