New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

FromKnowns consensus generator fails if no reads overlap a consensus #1362

Closed
fnothaft opened this Issue Jan 21, 2017 · 0 comments

Comments

Projects
None yet
1 participant
@fnothaft
Member

fnothaft commented Jan 21, 2017

This happens because we call a .min on an empty collection:

  org.apache.spark.SparkException: Job aborted due to stage failure: Task 0 in stage 8.0 failed 1 times, most recent failure: Lost task 0.0 in stage 8.0 (TID 28, localhost): java.lang.UnsupportedOperationException: empty.min
	at scala.collection.TraversableOnce$class.min(TraversableOnce.scala:209)
	at scala.collection.AbstractTraversable.min(Traversable.scala:105)
	at org.bdgenomics.adam.algorithms.consensus.ConsensusGeneratorFromKnowns.findConsensus(ConsensusGeneratorFromKnowns.scala:82)
	at org.bdgenomics.adam.rdd.read.realignment.RealignIndels$$anonfun$realignTargetGroup$1.apply(RealignIndels.scala:260)
	at org.bdgenomics.adam.rdd.read.realignment.RealignIndels$$anonfun$realignTargetGroup$1.apply(RealignIndels.scala:240)
	at scala.Option.fold(Option.scala:157)
	at org.apache.spark.rdd.Timer.time(Timer.scala:48)
	at org.bdgenomics.adam.rdd.read.realignment.RealignIndels.realignTargetGroup(RealignIndels.scala:240)
	at org.bdgenomics.adam.rdd.read.realignment.RealignIndels$$anonfun$realignIndels$1.apply(RealignIndels.scala:498)
	at org.bdgenomics.adam.rdd.read.realignment.RealignIndels$$anonfun$realignIndels$1.apply(RealignIndels.scala:498)

@fnothaft fnothaft added the bug label Jan 21, 2017

@fnothaft fnothaft added this to the 0.21.1 milestone Jan 21, 2017

@fnothaft fnothaft self-assigned this Jan 21, 2017

fnothaft added a commit to fnothaft/adam that referenced this issue Jan 23, 2017

[ADAM-1362] Fixing issue where FromKnowns consensus model fails if no…
… reads hit a target.

Fixes #1362:

* Adding ConsensusGeneratorFromKnownsSuite.
* ConsensusGeneratorFromKnowns drops any targets that did not get hit with any
  reads, instead of trying to identify the full reference region overlapped by
  reads. While this case (target with no reads) cannot happen in the FromReads
  consensus model, it can happen in the FromKnowns model (e.g., use all known
  INDELs from 1000G with a WES dataset).
* Additionally, tracked down a minor off by one bug in RealignIndels that caused
  one index in the consensus sequence to not get tested. This could cause reads
  that aligned to the end of the target to not get realigned.
* To improve legibility, broke up a set of long lines in RealignIndels.

@heuermh heuermh closed this in #1363 Jan 23, 2017

heuermh added a commit that referenced this issue Jan 23, 2017

[ADAM-1362] Fixing issue where FromKnowns consensus model fails if no…
… reads hit a target.

Fixes #1362:

* Adding ConsensusGeneratorFromKnownsSuite.
* ConsensusGeneratorFromKnowns drops any targets that did not get hit with any
  reads, instead of trying to identify the full reference region overlapped by
  reads. While this case (target with no reads) cannot happen in the FromReads
  consensus model, it can happen in the FromKnowns model (e.g., use all known
  INDELs from 1000G with a WES dataset).
* Additionally, tracked down a minor off by one bug in RealignIndels that caused
  one index in the consensus sequence to not get tested. This could cause reads
  that aligned to the end of the target to not get realigned.
* To improve legibility, broke up a set of long lines in RealignIndels.
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment