Chris:LabNotes/Moleculo/2013/2013-6-4
Jump to navigation
Jump to search
Quick Note on Celera Assembly error[edit]
Background[edit]
- Have been trying for the past few days to run Celera assembly on all Illumina read assembly + Moleculo, but have been encountering an error that would prevent the assembly to finish -> this is description of the solution
Solution to error[edit]
- Celera has been crashing with the following error report:
Fragment correction job 0001 failed. ================================================================================ runCA failed. ---------------------------------------- Stack trace: at /media/2TB_store1/cjwei/software/Hybrid_error_correction_assembly_software/w gs-7.0/Linux-amd64/bin/runCA line 1237 main::caFailure('1 overlap jobs failed; remove /media/disk-2/cjwei/13041 9_Onco...', undef) called at /media/2TB_store1/cjwei/software/Hybrid_error_corre ction_assembly_software/wgs-7.0/Linux-amd64/bin/runCA line 3923 main::overlapCorrection() called at /media/2TB_store1/cjwei/software/Hyb rid_error_correction_assembly_software/wgs-7.0/Linux-amd64/bin/runCA line 5880 ---------------------------------------- Failure message: 1 overlap jobs failed; remove /media/disk-2/cjwei/130419_Oncomelania_Illumina_Mo leculo/Triton_Hybrid_Assembly_5_13_2013/4_Illumina_All+Moleculo_Assembly-2/Illum ina_500bp+Moleculo_Assembly-2/3-overlapcorrection/frgcorr.sh to try again
- I figured out that the reason why this is is because the frgCorrBatchSize is too large by default so that memory would be exhausted. In order to fix this, there are two ways:
1) If the program has been running and do not want to restart, can go into the <3-overlapcorrection/frgcorr.sh> file and change...
jobid=`printf %04d $jobid` minid=`expr $jobid \* 200000 - 200000 + 1` maxid=`expr $jobid \* 200000` runid=$$
to...
jobid=`printf %04d $jobid` minid=`expr $jobid \* 20000 - 20000 + 1` maxid=`expr $jobid \* 20000` runid=$$
2) If starting the program, can include command in spec file and run celera as normally:
frgCorrBatchSize=2000 (or any integer <200000)
Plan to continue hybrid assembly[edit]
- Correct error and continue with Illumina_all_assembly+Moleculo_assembly
- Location of folder is in my genemapster: </media/disk-2/cjwei/130419_Oncomelania_Illumina_Moleculo/Triton_Hybrid_Assembly_5_13_2013/4_Illumina_All+Moleculo_Assembly>