Re: [wgs-assembler-users] Consensus and extendClearRanges failure

[Walenz, B. <bw...@jc...>]

Hi, Xueping-

The consensus ‘errors’ aren’t a problem.  They’re detected and resolved by utgcnsfix.  The logs are in 5-consensus/consensus-fix.out if you want to see them.  We need to do a bit of work on the pipeline to clean this up, but its low priority.

The ECR failure has also been reported in CGW.  We’ll need an example (and some time) to figure out what’s going wrong.  The fix on the wikipage is a little out of date.  A simpler method is to skip the entire scaffold, or the gap, that is causing the problem.  In the shell script, add either:

-s <scaffold-id>
-S <gap-number>

to the extendClearRanges command.  If you can figure out which gap number it is, use that, otherwise skip the whole scaffold.  Gap numbers will change based on the –b/-e scaffold ranges — meaning you can’t run just this scaffold to test which gap is failing.

Can you send the full log (to me, no need to send to the list)?

b


FYI,

usage: extendClearRanges [opts] -c ckpName -n ckpNumber -g gkpStore

  -c ckpName     Use ckpName as the checkpoint name
  -n ckpNumber   The checkpoint to use
  -g gkpStore    The gatekeeper store

  -C gap#        Start at a specific gap number
  -b scafBeg     Begin at a specific scaffold
  -e scafEnd     End after a specific scaffold (INCLUSIVE)

  -o scafIID     Process only this scaffold
  -s scafIID     Skip this scaffold
  -O gap#        Process only this gap
  -S gap#        Skip this gap

  -i iterNum     The iteration of ECR; either 1 or 2

  -load          Load gkpStore into memory

  -V             Enable VERBOSE_MULTIALIGN for debugging




On 11/7/12 4:50 AM, "Quan, Xueping" <x....@im...> wrote:

Dear All

My hybrid assembly (with about 131GB Illumina reads and 2GB 454 reads).  Now it failed at the extendClearRanges stage. But when I checked all the stage. I found that in the 5-consensus/ folder, there are 110 jobs, only 60 of them succeed (with both err and success output file). Below is the last few lines of the job61.err
"NumColumnsInUnitigs             = 261474722
NumGapsInUnitigs                = 280587
NumRunsOfGapsInUnitigReads      = 11234803
NumColumnsInContigs             = 0
NumGapsInContigs                = 0
NumRunsOfGapsInContigReads      = 0
NumAAMismatches                 = 0
NumVARRecords                   = 0
NumVARStringsWithFlankingGaps   = 0
NumUnitigRetrySuccess           = 0

WARNING:  Total number of unitig failures = 1

Consensus did NOT finish successfully."

Should I re-run the consensus and how can to make the failed consensus jobs succeed?

The failure information in file extendClearRanges-scaffold.0017585.err is
"FRG type R ident 757459763 container  94540021 parent  94540021 hanginsertMultiAlign()-- ERROR: multialign 636535 C has invalid fragment/unitig layout (lengt
h 2214) -- exceeds bounds of consensus sequence (length 2215).
extendClearRanges: MultiAlignStore.C:327: void MultiAlignStore::insertMultiAlign(MultiAlignT*, bool, bool): Assertion `GetMultiAlignLength(ma) <= GetMultiAli
gnLength(ma, true)' failed.
     57      0 position   2106   2185
FRG type R ident 757459764 container  94540021 parent  94540021 hang     57      0 position   2106   2185
FRG type R ident 930433904 container  94540021 parent  94540021 hang     61      0 position   2185   2110
UTG type X ident    636535 position      0   2214 num_instances 0"

Can I follow this instruction to restart it
http://sourceforge.net/apps/mediawiki/wgs-assembler/index.php?title=Extend_clear_ranges_failure?

Thanks very much!






Dr. Xueping Quan

Imperial College London