pacbioToCA failed

[shiningway]

I’m trying to use pacbioToCa to correct the pacbio reads with some 90bp pair end Illumina reads. This a 11Mb bacterial genome. It failed at the overlap.sh step. There were 49 overlap.sh processes and only 4 of them generated output files. I'm using a stand alone Linux machine with 8 core and 16GB memory.

Some out files in the 0-overlaptrim-overlap directory show this kind of err:
~/0-overlaptrim-overlap/overlap.sh: line 61: 9271 Killed
$bin/overlapInCore -G --hashbits 24 --hashload 0.75 -t 2 $opt
-k 10 -k /data1/PacBioClip30x/0-mercounts/asm.nmers.obt.fasta -o /data1/PacBioClip30x/0-overlaptrim-overlap/$bat/$job.ovb.WORKING.gz -H 1-1 -R 1-1 /data1/PacBioClip30x/asm.gkpStore

Some other out files in the same directory show memory issue:
Could not malloc memory (3623878656 bytes)
overlapInCore: AS_UTL_alloc.C:73: void* safe_malloc(size_t): Assertion `p != __null' failed.


The spec file I used is the following. Could you tell me what I can add/modify to run pacbioToCA? Thanks a lot for any help.

# original asm settings
utgErrorRate = 0.25
utgErrorLimit = 6.5

cnsErrorRate = 0.25
cgwErrorRate = 0.25
ovlErrorRate = 0.25

#merSize=14

merylMemory = 128000
merylThreads = 16

ovlStoreMemory = 8192

# grid info
useGrid = 0
scriptOnGrid = 0
frgCorrOnGrid = 0
ovlCorrOnGrid = 0

ovlHashBits = 24
ovlThreads = 2
ovlHashBlockLength = 20000000
ovlRefBlockSize = 50000000

frgCorrThreads = 2
frgCorrBatchSize = 100000

ovlCorrBatchSize = 100000

ovlConcurrency = 6
cnsConcurrency = 16
frgCorrConcurrency = 8
ovlCorrConcurrency = 16
cnsConcurrency = 16

#shortReads
frgMinLen = 65 # smaller than your read length
ovlMinLen = 48 # about 75% of your frgMinLen
merSize = 10

[jbingham]

From the pacBioToCA wiki page:

"If you have Illumina sequences longer than 64bp but shorter than 100bp, have to specify the

-shortReads

option to the pipeline"

Not sure if that's the issue, or part of it. pacBioToCA works best with reads at least 100 bp long.

Also, if you have enough PacBio coverage, you can check out HGAP. This is how PacBio has gotten single contig assemblies with bacteria. It gets better assemblies without using the Illumina reads at all -- provided you have the coverage.

https://github.com/PacificBioscience...-Process-(HGAP)

[shiningway]

From the pacBioToCA wiki page:

"If you have Illumina sequences longer than 64bp but shorter than 100bp, have to specify the

-shortReads

option to the pipeline"

Thanks. I looked at that too. And I put '-shortReads' in pacbio.spec file. But it did not work. runCA does not recognize -shortReads option.

Is HGAP part of pacbio's smrtanalysis-1.4.0? I'm trying that.

[lhon]

runCA should only be used after error correction to assemble corrected reads. The -shortreads option is part of pacBioToCA.

HGAP is also part of SMRT Analysis 1.4, and details around that implementation is documented here:

https://github.com/PacificBiosciences/Bioinformatics-Training/wiki/HGAP